This subpage provides a detailed overview of all the freelance services that I prefer to do right now. Keep in mind: the list is not exhaustive, and we may have a deal on things that goes beyond it.

Last updated on Apr 5, 2025.

1

Modular Python software

Modular software is built with reusable, well-structured components, making it scalable and easy to maintain. This service is ideal for individuals or teams needing complex and organized programming projects. Whether you're building a tool, user application or library, modular design ensures flexibility and long-term reliability.

What you get

• A Python codebase organized into modules and packages
• Features like clean APIs, clear documentation and error handling
• Scalable design, enabling future enhancements and integrations
• Fully tested software with unit and integration tests included

Why you might need it

• Save development time: reuse modules for future projects or extensions
• Improve collaboration: structured code makes teamwork and onboarding smoother
• Ensure reliability: robust, tested modules minimize bugs and downtime

What I do

• Plan the software architecture based on your project requirements
• Develop Python modules with clear input/output
• Use best practices for structuring codebases (e.g., following PEP-8 guidelines)
• Test modules using Python testing frameworks (Pytest, Unittest)
• Deliver the software with clear setup instructions, module documentation and API documentation

Examples

👁️Kallisto       

LLM-based OSINT tool designed to perform deep web searches by orchestrating multiple web agents and a knowledge agent that uses state-of-the-art machine learning methods. The tool crawls the entire web to gather massive amounts of publicly available data, then leverages advanced LLM techniques to perform natural language tasks on that information.

👁️ Kallisto    — LLM-based OSINT tool designed to perform deep web searches by orchestrating multiple web agents and a knowledge agent that uses state-of-the-art machine learning methods. The tool crawls the entire web to gather massive amounts of publicly available data, then leverages advanced LLM techniques to perform natural language tasks on that information.

🤔Pochemuchka       

Automatic prompt engineering, testing and load balancing for AI models in production and R&D

🤔 Pochemuchka    — Automatic prompt engineering, testing and load balancing for AI models in production and R&D

Prerequisites

• A description of the project, its goals and the functionality required
• Any existing code or systems to integrate with
• Specifications for APIs, libraries or frameworks to be included
• (Optional) Examples of desired architecture or project templates

Preliminary pricing

Highly depends on features and scope, but ~$200 is the average price for complex multi-module applications. Large-scale systems (e.g., enterprise-grade software with robust documentation) may cost $400–$700.

Payment model

Payment is normally pre-fixed and always per-deliverable.

2

Web scraping

Web scraping involves extracting valuable data from websites in a structured format. If you're a business owner, marketer, researcher or data analyst, this service is for you. Need to track prices, monitor trends, collect leads or gather bulk data for research? Here's an efficient way to acquire information that can help you.

What you get

• A clean, well-organized dataset in your preferred format (Excel, CSV, JSON or database-ready)
• Screenshots and sample outputs to ensure you're receiving exactly what you need
• (Optional) Scripts used to collect the data

Why you might need it

• Save money: no more manual data collection, which can be time-consuming and error-prone
• Create revenue opportunities: leverage market data to make better business decisions
• Stay ahead of competitors: gain insights into trends, customer preferences and more

What I do

• Analyze your data requirements and target websites
• Use scraping & HTTP Python tools (BeautifulSoup, Scrapy, requests) or Selenium to create scripts
• Develop custom solutions that bypass restrictions like CAPTCHA or dynamic loading
• Automatically clean and preprocess the extracted data inside scripts using Pandas pipelines (if needed)
• Deliver the final dataset in the format of your choice

Examples

🫐Wildberries parsers       

Python scripts for parsing Wildberries, the largest Russian online retailer. Designed to extract data via API.

🫐 Wildberries parsers    — Python scripts for parsing Wildberries, the largest Russian online retailer. Designed to extract data via API.

Prerequisites

• The URLs of target websites and a clear description of the type of data you need to collect
• Information on the desired format of the output data

Preliminary pricing

The average price for scraping static pages with minimal formatting is ~$40. Handling pagination, login authentication or dynamic content may cost $60–$100. More advanced features (e.g., bypassing captchas, integrating proxies or scraping multiple sites with API development) are highly unpredictable in price.

Payment model

Payment is always per-deliverable and pre-fixed.

3

Data processing

Data processing turns raw data into the required form. Whether you're dealing with messy spreadsheets, incomplete databases or large volumes of unstructured information, this service helps you make sense of it. Ideal for small businesses, analysts or researchers who need to clean, organize and visualize data for decision-making.

What you get

• A structured and cleaned dataset ready for analysis
• (Optional) Scripts used to process the dataset

Why you might need it

• Save time: no more manual data cleaning or repetitive tasks
• Streamlined operations: enable efficient decision-making with organized and accurate data

What I do

• Depending on the source, load the data into my workspace to use Python tools (Pandas, NumPy), SQL or Google Sheets (Excel)
• Perform data validation
• Clean, sort and standardize raw data
• Handle missing values, inconsistencies and duplicates
• Perform advanced data transformation, like pivoting, merging or aggregating

Examples

🔎Paysage       

Pandas add-on library: find data quality issues, clean/improve dataframes and generate reports — in one line, using scikit-learn transformer

🔎 Paysage    — Pandas add-on library: find data quality issues, clean/improve dataframes and generate reports — in one line, using scikit-learn transformer

⚡User activity analytics       

End‑to‑end real-time analytics pipeline using PySpark Structured Streaming. The pipeline ingests user activity logs from Kafka, applies multiple levels of transformations and aggregations, and computes both basic and session-level metrics. It includes dynamic anomaly detection using moving averages and standard deviations, and flags unusual spikes in activity.

⚡ User activity analytics    — End‑to‑end real-time analytics pipeline using PySpark Structured Streaming. The pipeline ingests user activity logs from Kafka, applies multiple levels of transformations and aggregations, and computes both basic and session-level metrics. It includes dynamic anomaly detection using moving averages and standard deviations, and flags unusual spikes in activity.

Prerequisites

• The raw data files or access to your data source
• Details on the specific processing tasks or desired output
• (Optional) Information about the business problem or question you're trying to solve with the data

Preliminary pricing

$40–$100 for basic tasks like cleaning datasets or simple transformations, depending on the amount of data. Handling large datasets with big data tools or creating reusable workflows may cost $100–$200. Advanced workflows (real-time processing, integrating pipelines) or working with specialized formats may cost more. The average price for this service is ~$70.

Payment model

In some cases, payment may be on an hourly basis.

4

Data visualization

Data visualization turns numbers into understandable forms that can be presented to non-technical people or simply to see something important in those numbers. This service is perfect for business leaders, marketers, analysts or anyone looking to communicate insights clearly and effectively.

What you get

• Professionally designed charts, graphs, maps or infographics based on your data
• (Optional) Different variations of static visuals for presentations or reports

Why you might need it

• Better communication: simplify complex data and understand its nature
• Increased engagement: captivate your audience with visually striking graphics
• Improved decision-making: quickly identify trends, anomalies and opportunities

What I do

• Analyze your data and determine the most effective visualization techniques
• Design visuals using Python tools (Matplotlib, Seaborn, Plotly and others) or Superset/Tableau
• Customize colors, layouts and interactivity to align with your branding or preferences
• Deliver exportable visuals compatible with your platforms
• Provide guidance on interpreting and using the visuals effectively

Examples

📊Telegram chat analyzer       

Streamlit-based interface for uploading Telegram chat exports and obtaining detailed statistics/visualizations using various data analysis techniques and natural language processing. Also supports chat exports from WhatsApp.

📊 Telegram chat analyzer    — Streamlit-based interface for uploading Telegram chat exports and obtaining detailed statistics/visualizations using various data analysis techniques and natural language processing. Also supports chat exports from WhatsApp.

Prerequisites

• A dataset or access to your data source
• Information on what you want to showcase or the story you want to tell
• Preferred platforms or tools if applicable
• (Optional) Branding guidelines or example visuals you admire

Preliminary pricing

$30–$50 for basic charts on common data types in Python, Excel or Google Sheets. $80–$200 for interactive/animated visualizations and uncommon data types. $200+ for highly customized and detailed visuals, advanced interactivity or visualizations on multiple datasets. These prices are mostly determined by urgency.

Payment model

In some cases, payment may be on an hourly basis.

5

Exploratory data analysis

Exploratory data analysis uncovers patterns, relationships and anomalies in your data. Whether you're preparing for a data science project, trying to understand your customers or optimizing operations, EDA provides insights and ensures your data is ready for advanced analytics. This service is essentially a combination of data visualization and statistical analysis.

What you get

• An in-depth overview of your data, including key statistics, distributions and correlations
• Visualizations to explore trends and relationships
• A concise report summarizing findings, potential insights and recommendations
• (Optional) Data cleaning and preparation to remove outliers, handle missing values and standardize formats

Why you might need it

• Deeper insights: reveal hidden trends and patterns in your data
• Data quality assurance: ensure your data is clean and reliable for further analysis
• Informed decisions: gain a solid foundation for strategy and predictive modeling

What I do

• Depending on the source, load the data into my workspace to use Python tools, SQL or Google Sheets (Excel)
• Perform data cleaning (Pandas, NumPy)
• Identify outliers, missing values and potential biases in the dataset
• Prepare the data for further modeling or analysis by transforming and normalizing it
• Conduct descriptive statistical analysis to summarize your data
• Visualize relationships and distributions (Seaborn, Matplotlib)
• Make a report (.pdf, Juputer Notebook, Google Colab, etc.)

Examples

Prerequisites

• The dataset or access to your data source
• A clear understanding of your goals (e.g., uncover trends, assess data quality or identify potential issues)
• Preferred output formats
• (Optional) A description of the specific problem you're addressing or the decisions you aim to inform with the findings

Preliminary pricing

$50–$150 for summary statistics, correlations and basic charts with explanations. $150–$250 for in-depth analysis, probably using feature engineering or specific problem-focused methods. Comprehensive reports on multiple large datasets or with advanced visualizations may vary in price highly.

Payment model

Payment is often on an hourly basis.

6

Fine-tuning

ML model fine-tuning involves optimizing pre-trained models to meet specific business needs. This service is ideal for businesses and researchers who want to maximize the performance of existing models for tasks like image classification, natural language processing or predictive analytics.

What you get

• A fine-tuned ML model ready to solve your specific problem
• Comprehensive evaluation metrics to validate performance
• A detailed report explaining the modifications, performance improvements and potential applications
• (Optional) Deployment-ready code for integration into your existing systems

Why you might need it

• Save time and resources: use existing powerful models without starting from scratch
• Customized solutions: tailor models to your unique dataset and objectives
• Enhanced accuracy: improve predictions, classifications or forecasts to drive better results

What I do

• Analyze your dataset, processses and objectives to identify the most suitable pre-trained model
• Fine-tune a baseline model using TensorFlow/PyTorch frameworks
• Optimize hyperparameters
• Perform rigorous evaluation using your data and refine the model iteratively
• (Optionally) Perform A/B testing of the model
• Deliver/integrate the fine-tuned model with clear documentation and guidance for use
• (Optionally) Monitor the model performance metrics in production for potential optimization

Examples

🪶MobileEAST       

My thesis project: the lightweight and fast scene text detection based on EAST architecture and MobileNet layers

🪶 MobileEAST    — My thesis project: the lightweight and fast scene text detection based on EAST architecture and MobileNet layers

Prerequisites

• A labeled dataset or clear instructions on where to source training data
• Details about the task you want to solve
• Information on the deployment environment (e.g., cloud platform, local servers)
• (Optional) Access to computational resources for model training

Preliminary pricing

$100–$400 for tuning pre-trained models with perfomance validation, depending on many factors such as quality of data, problem scope and domain.

Payment model

Payment is normally pre-fixed and always per-deliverable.

7

Frontend

Frontend development focuses on creating engaging, user-friendly web applications. If you need a visually appealing interface, this service brings your ideas to life.

What you get

• A responsive, modern frontend for your needs
• Smooth navigation, interactive elements and an optimized UX
• Cross-browser compatibility and mobile-friendly designs
• Clean, well-documented code for seamless future updates or scaling

Why you might need it

• Enhanced user engagement: create interfaces that are easy and enjoyable to use
• Increased conversions: guide users effectively to achieve your business goals

What I do

• Design and develop the frontend using React framework
• Write clean, maintainable JavaScript/TypeScript components and CSS modules
• Implement animations, transitions and responsive design for a dynamic experience
• Test for compatibility across browsers and devices

Examples

🚗Traffic Observer       

A complete road traffic analysis system for roundabouts: real-time processing from multiple cameras, integration with Kafka for messaging, InfluxDB for time series storage and Grafana for interactive dashboards. Processes RTSP streams or local .mp4 files, computes per-road vehicle count and intensity (vehicles per minute), supports parallel processing.

🚗 Traffic Observer    — A complete road traffic analysis system for roundabouts: real-time processing from multiple cameras, integration with Kafka for messaging, InfluxDB for time series storage and Grafana for interactive dashboards. Processes RTSP streams or local .mp4 files, computes per-road vehicle count and intensity (vehicles per minute), supports parallel processing.

👂Minimal voice assistant       

An implementation of a simple extensible voice-controlled assistant with Node backend and React frontend

👂 Minimal voice assistant    — An implementation of a simple extensible voice-controlled assistant with Node backend and React frontend

Prerequisites

• A detailed description of your project goals and target audience
• (Optional) Wireframes, mockups or design inspiration
• (Optional) Branding guidelines (logos, colors, fonts) for consistency
• (Optional) Access to APIs or backend services if the frontend will integrate with them

Preliminary pricing

~$150 for frontend interfaces or single-page apps of moderate complexity with interactive elements and responsive designs. API integration and additional React modules may increase the price.

Payment model

Payment is always per-deliverable, but the price isn't guaranteed to be pre-fixed and depends on the client's revisions.

8

Static websites

Static websites are lightweight, fast-loading websites that deliver consistent content to users. Perfect for personal portfolios, landing pages, blogs or event sites, static websites are cost-effective and require minimal maintenance. If you want a fast, secure and SEO-friendly online presence, this service is for you. Look, I even made one for myself.

What you get

• A responsive website for your needs
• Up to 10 pages (e.g., Home, About, Services, Contact) designed to showcase your content effectively
• Optimized for speed, security and search engines (SEO)
• Hosted and deployed on platforms like GitHub Pages, Netlify or Vercel

Why you might need it

• Save costs: no backend systems mean lower hosting fees and maintenance needs
• Faster load times: delight users with a seamless browsing experience
• Enhanced security: static sites are less prone to hacking compared to dynamic ones

What I do

• Develop the pages and components using React framework (JavaScript/TypeScript)
• Implement primitive Next.js backend or use SSG (Gatsby & GraphQL, Jekyll)
• Incorporate tools like Tailwind CSS for responsive design
• Optimize images, fonts and code to ensure fast load times
• Implement SEO best practices like metadata, alt text and site maps
• Host the website

Examples

🚀avrtt.blog       

Static MDX-based (and simply based) website with SEO — designed and written completely from scratch & with love using React, Gatsby, TypeScript and Sass modules

🚀 avrtt.blog    — Static MDX-based (and simply based) website with SEO — designed and written completely from scratch & with love using React, Gatsby, TypeScript and Sass modules

🏠Next.js website template       

A simply designed and catchy website that can be used as a blog, contact page and portfolio, built with React, Next.js & Tailwind CSS

🏠 Next.js website template    — A simply designed and catchy website that can be used as a blog, contact page and portfolio, built with React, Next.js & Tailwind CSS

Prerequisites

• A clear description of the website's purpose and target audience
• Content for the website (e.g., text, images, logos)
• Design preferences or examples of websites you admire
• (Optional) A domain name and/or hosting preference

Preliminary pricing

$150–$250 for basic 2-3 pages website (personal portfolio, small business pages). ~$400 for multi-page websites with contact forms, SEO and animations. Extensive custom design and JavaScript functionality may vary these prices significally.

Payment model

Payment is always per-deliverable, but the price isn't guaranteed to be pre-fixed and depends on the client's revisions.

9

Bots

Bots automate repetitive tasks and improve efficiency across various applications. This service is for everyone, especially businesses, marketers and individuals looking to save time and enhance productivity.

Here are the most popular types of bots I've worked on:

• Informational bot: collects/aggregates any data and notifies the user when needed
• Assistant: automates the work of a consultant; allows you to guide the user and answer frequently asked questions according to a consistent algorithm
• Chatbot: allows you to bring the functionality of popular language models to other platforms to create conversations; can be modified and trained on your data in order to answer non-standard questions from customers regarding your product, service, etc.
• Auto commenter: a userbot that monitors new posts in specified channels/groups/chats and instantly sends AI-generated comments below; each comment is generated based on the text of the post and your pre-defined bot behavior (prompt)

It's hard to classify bots clearly. Usually, they utilize a variety of different tools and methods.

What you get

• A customized bot designed to handle specific tasks
• Fully tested and deployed to your preferred platform (e.g., Telegram, Instagram, Discord, your website, standalone scripts)
• (Optional) Clear documentation on how to operate and maintain the bot

Why you might need it

• Save time: automate repetitive tasks, freeing up valuable resources
• Improve engagement: chatbots provide 24/7 customer support or lead generation
• Streamline operations: automate data entry, notifications or even advanced tasks like workflow orchestration

What I do

• Assess your requirements and choose the right technologies
• Develop the main bot components (Python)
• Integrate language model functionality
• Test and refine the bot for accuracy, speed and reliability
• Integrate the bot with your systems, platforms or APIs as needed

Examples

🤖Auto-comment bot (Telegram)       

Software with Telegram bot-like functionality that allows to leave instant GPT4-generated comments by specifying Telegram channels to monitor for new posts, and prompts to customize behavior

🤖 Auto-comment bot (Telegram)    — Software with Telegram bot-like functionality that allows to leave instant GPT4-generated comments by specifying Telegram channels to monitor for new posts, and prompts to customize behavior

👮Telegram Content Moderator       

A Telegram bot with NLP and vision transformer that detects and moderate messages in real-time, logging actions and auto-removing/warning users. Designed for Telegram groups.

👮 Telegram Content Moderator    — A Telegram bot with NLP and vision transformer that detects and moderate messages in real-time, logging actions and auto-removing/warning users. Designed for Telegram groups.

Prerequisites

• A clear description of the task or problem you want the bot to address
• Access to the data sources, APIs, bot accounts or platforms the bot will interact with
• (Optional) Information on the preferred bot hosting or deployment platform (if applicable)
• (Optional) Examples or specifications for desired bot behavior

If you order an userbot, then make sure there's nothing confidential or important on it. To access accounts API, in Telegram for instance, I need phone numbers with confirmation codes to log in, and passwords if you have two-factor authentication.

Preliminary pricing

This service is highly unpredictable in price. In average, an NLP-driven bot of medium complexity with APIs integration or databases costs ~$130.

Payment model

Payment is normally pre-fixed and always per-deliverable.

10

Data collection software

Data collection software automates the gathering of structured or unstructured data from various sources like websites, APIs or databases. If you need to collect dynamic data for research, analytics, reporting or decision-making, this service will provide you with a reusable tool.

What you get

• Custom-built software for your data collection needs
• Data delivered in a structured format like Excel, CSV or a database
• Features like error handling, scheduling and automated updates
• Documentation and training on how to use the software effectively

Why you might need it

• Save time: automate tedious data collection tasks
• Improve decision-making: access high-quality data to power analytics and insights
• Stay competitive: gather real-time data for trends, pricing or customer behavior

What I do

• Analyze your data collection requirements
• Develop software using Python tools (BeautifulSoup, Scrapy, requests) and API integrations
• Test the software extensively to ensure accuracy and reliability
• Deliver the tool with documentation

Examples

🫐Wildberries parsers       

Python scripts for parsing Wildberries, the largest Russian online retailer. Designed to extract data via API.

🫐 Wildberries parsers    — Python scripts for parsing Wildberries, the largest Russian online retailer. Designed to extract data via API.

⛏️training-scraper       

A Python script designed for web scraping, data integration and further model training. It leverages BeautifulSoup for parsing HTML content, TensorFlow and Keras for building and training baseline models, and several other libraries for data processing and automation.

⛏️ training-scraper    — A Python script designed for web scraping, data integration and further model training. It leverages BeautifulSoup for parsing HTML content, TensorFlow and Keras for building and training baseline models, and several other libraries for data processing and automation.

Prerequisites

• A clear description of the data you need and where it can be sourced from
• Information on the preferred output format or storage location
• (Optional) Access to APIs, databases or websites requiring authentication

Preliminary pricing

$100–$150 for one-time data extraction tool. $150–$300 for tools with scheduling, filtering or error-handling. Real-time data collection may double the price.

Payment model

Payment is normally pre-fixed and always per-deliverable.

11

Hypothesis testing

Hypothesis testing is a statistical method for evaluating assumptions about data. It's perfect for researchers, analysts and business owners who need to make data-driven decisions (by testing marketing strategies, customer preferences, product effectiveness and so on).

What you get

• Statistical analysis performed using industry-standard techniques
• Report that explain the results in a simple way
• Recommendations or next steps based on the findings

Why you might need it

• Validate ideas: make confident decisions based on data, not guesses
• Improve strategies: identify what works and what doesn't for your goals
• Optimize resources: focus efforts on initiatives with proven success potential

What I do

• Define the null and alternative hypotheses based on your objectives
• Collect or clean data to ensure it meets testing requirements
• Perform statistical tests using Python tools (SciPy, Statsmodels and many others)
• Create clear visuals (e.g., graphs, p-value tables) to interpret the results
• Write a detailed report summarizing the findings and their implications

Examples

⚒️ABforge       

A feature-rich library for running Bayesian A/B tests for experiments and key metrics analysis

ABforge is designed for a variety of use cases, utilizing the Bayesian approach to implement different variables and tests, including binary, Poisson, normal, delta-lognormal and discrete tests on different metrics. It's especially useful for analysing key metrics in marketplaces, such as conversion rates, ticket size, ARPU differences between test variants, etc. The main idea of this library is to create an all-in-one tool for running A/B tests separately from closed-source code and business logic.

⚒️ ABforge    — A feature-rich library for running Bayesian A/B tests for experiments and key metrics analysis

Prerequisites

• A setup and explanation of the hypothesis to be tested, or a research question
• A dataset relevant to the hypothesis (raw or cleaned)
• Information on the context of the problem (e.g., what decision will the results impact?)
• (Optional) Benchmarks or expected results for comparison

Preliminary pricing

This service depends on many factors, price is negotiable.

Payment model

Payment is normally on an hourly basis.

12

Dashboards

Dashboards are interactive visual tools that aggregate, analyze and display data in real time. Ideal for business leaders, analysts and teams needing to monitor KPIs, track performance or derive insights quickly.

What you get

• A custom-built, interactive dashboard
• Integration with data sources (Excel, databases, APIs, etc.) for real-time updates
• Clean, user-friendly layouts designed to display critical metrics effectively
• (Optional) Tutorials or guides to help your team navigate and utilize the dashboard

Why you might need it

• Improved decision-making: access insights instantly
• Time savings: consolidate data from multiple sources into one view
• Enhanced collaboration: share insights seamlessly with your team or stakeholders

What I do

• Understand your key metrics, pick the right tool and design a layout that meets your needs
• Build a dashboard using the chosen tool (see below)
• Connect the dashboard to your data sources for live or automated updates
• Optionally use helper Python libraries (Matplotlib, Seaborn, Pandas) and SQL
• Optimize visuals for clarity, ensuring metrics are easy to interpret
• Provide training or documentation for effective use and updates

How I usually choose a tool to create a dashboard:

• Google Sheets and/or Looker Studio (alt. Excel): for simple, ad-hoc projects (small-scale, highly portable dashboards, mostly for non-technical clients) and projects requiring little scripting (Google Apps Script); usually GS acts as a lightweight backend for Looker Studio (formerly Google Data Studio)
• Dash: for complex, interactive and highly customizable dashboards, especially for embedding analytics in web apps or other code-driven projects (it's a part of Plotly, a Python library)
• Apache Superset: for projects that rely heavily on SQL and/or as a lightweight & open-source alternative of Tableau (it's no-code, therefore it's faster)
• Grafana: for real-time monitoring projects, especially when working with time series; probably the best option for DevOps working with cloud systems like Prometheus, Elasticsearch, etc.
• Tableau Public: for simple dashboards, fast building (it's no-code too), demos and quick sharing with non-technical clients

Besides these tools, some “exotic” types of dashboards can be created using Streamlit. Streamlit apps development is moved to a separate type of services.

Examples

🔁dashbordina       

Marketing analytics dashboard for audience segmentation and multi-channel ROI analysis with ETL, implemented with Dash and Airflow

🔁 dashbordina    — Marketing analytics dashboard for audience segmentation and multi-channel ROI analysis with ETL, implemented with Dash and Airflow

♟️superset-supremacy       

Apache Superset dashboard prototype based on strategic marketing project (channel prioritization, budget allocation, ROI forecasting)

♟️ superset-supremacy    — Apache Superset dashboard prototype based on strategic marketing project (channel prioritization, budget allocation, ROI forecasting)

Prerequisites

• A clear understanding of the KPIs or data you want to track
• Access to your data sources (e.g., files, databases, APIs)
• (Optional) Branding guidelines or design preferences
• (Optional) Examples of dashboards you like or features you need (e.g., filters, drill-downs)
• (Optional) Preferred dashboard development tool from the list above

Preliminary pricing

Dashboards with simple visualizations and KPIs cost $100–$300. Custom features with advanced analytics, real-time updates and handling many data sources may cost $300–$800. The average price is ~$350.

Payment model

Payment is normally on an hourly basis.

13

Technical writing

Technical writing transforms complex concepts into clear, concise documentation. This service is perfect for businesses needing user manuals, API documentation, technical blogs or guides. Whether you're targeting developers, users or potential clients, you get materials that are professional and easy to understand.

What you get

• Well-organized, professional text to your audience in preferred writing style
• Accurate and concise language, free from ambiguity
• (Optional) Visual aids such as flowcharts, diagrams or screenshots

Why you might need it

• Streamlined communication: help users or teams understand your tools and systems
• Increased efficiency: deduce support requests with clear, self-service documentation
• Enhanced credibility: professional materials reflect the quality of your product or service

What I do

• Research and analyze your product or process to understand its functionality
• Organize and structure content based on user needs and objectives
• Write and edit clear, accurate and engaging technical documents
• Ensure consistency with your brand's tone and style

Examples

🚀The Research section (avrtt.blog)       

My tech blog about all the cool things I research along the career path within data science

🚀 The Research section (avrtt.blog)  — My tech blog about all the cool things I research along the career path within data science

🤓avrrt.course       

My free course on machine learning and data science

🤓 avrrt.course  — My free course on machine learning and data science

Prerequisites

• A description of the product, process or system needing documentation
• Access to relevant materials, such as prototypes, systems or existing notes
• Details about your target audience and their familiarity with the subject
• (Optional) Examples of documentation styles or formats you prefer

Preliminary pricing

Simple documentation cost $50–$150 (guides or FAQs), depending on text length. API documentation and user manuals cost $150–$300. Complex technical whitepapers may cost more.

Payment model

Payment is always on an hourly basis.

14

Scripting

Scripting involves writing code to automate tasks, enhance workflows or solve specific problems. If you're looking to save time or add custom functionality to your tools, this service can help you.

What you get

• Custom scripts to automate tasks, process data or solve specific challenges
• Fully tested, efficient and error-free code in your preferred scripting language (Python, Bash, PowerShell)
• Clear instructions for running the scripts and integrating them into your systems

Why you might need it

• Save time: automate repetitive tasks, freeing up resources
• Improve accuracy: minimize errors in manual processes
• Increase productivity: speed up workflows with efficient, reusable scripts

What I do

• Analyze your needs and the tasks to be automated or solved
• Develop Python scripts and/or Shell scripts
• Test the scripts extensively
• Deliver the scripts with usage instructions

Examples

✉️mailman       

Personalized, LLM-based email marketing automation with user segmentation and A/B testing analytics

✉️ mailman    — Personalized, LLM-based email marketing automation with user segmentation and A/B testing analytics

🤔Pochemuchka       

Automatic prompt engineering, testing and load balancing for AI models in production and R&D

🤔 Pochemuchka    — Automatic prompt engineering, testing and load balancing for AI models in production and R&D

Prerequisites

• A clear description of the task or workflow you want to automate or enhance
• Access to the systems, data or tools the script will interact with
• Preferred scripting language (if applicable)
• (Optional) Examples or specifications for desired script functionality

Preliminary pricing

$50–$150 for simple automation tasks, $150–$400 for handling data processing, working with databases or API integration. $500–$1000+ for end-to-end workflow or highly complex logic. The averace price is ~$80.

Payment model

Payment is always per-deliverable and pre-fixed.

15

Streamlit apps

Streamlit apps turn your data, models or algorithms into interactive web applications. Ideal for data scientists, ML engineers and businesses wanting to share insights, predictive tools or custom solutions quickly, without the need for a full website.

What you get

• A sleek, interactive web application built using Streamlit
• Features like dropdowns, sliders, charts and tables to enhance user interaction
• Seamless integration with your data sources, APIs or machine learning models
• Fully deployable and accessible on the web or internal platforms

Why you might need it

• Showcase insights: share interactive reports with your team or clients
• Streamline workflows: provide an easy-to-use interface for complex processes or models
• Save time: rapidly build and deploy apps without extensive coding overhead

What I do

• Develop the application using Streamlit, integrating necessary libraries
• Add features like file upload, live data updates and interactive widgets
• Test the app (usability, performance, reliability)
• Assist with deployment on platforms

Examples

📊Telegram chat analyzer       

Streamlit-based interface for uploading Telegram chat exports and obtaining detailed statistics/visualizations using various data analysis techniques and natural language processing. Also supports chat exports from WhatsApp.

📊 Telegram chat analyzer    — Streamlit-based interface for uploading Telegram chat exports and obtaining detailed statistics/visualizations using various data analysis techniques and natural language processing. Also supports chat exports from WhatsApp.

🐋QASATIK       

LLM-based Q&A on preloaded docs, raw data, Wikipedia articles and scraped web pages with knowledge graphs, analytics, charts and Streamlit interface

QASATIK is an app dedicated to helping interrogate large volumes of documents, data files and web pages. Built with Streamlit, it supports file uploads, online article scraping and querying using configurable language models (OpenAI, LangChain and LlamaIndex). In addition, it provides interactive knowledge graph visualizations, analytics and charting utilities to help explore and understand the data.

🐋 QASATIK    — LLM-based Q&A on preloaded docs, raw data, Wikipedia articles and scraped web pages with knowledge graphs, analytics, charts and Streamlit interface

Prerequisites

• A clear description of the app's purpose and target audience
• Access to data, machine learning models or APIs the app will utilize
• Preferred hosting platform for deployment
• (Optional) Examples of similar apps or specific features you want included

Preliminary pricing

Highly depends on initial software. The average price for static data displays and basic interactivity is ~$100.

Payment model

Payment is normally pre-fixed and always per-deliverable.

16

Academic help

This service provides expert assistance with coding projects, Linux tasks, probability theory, statistics or research-related coursework and thesis requirements. Perfect for students, researchers and professionals aiming to enhance their understanding, complete assignments, or achieve academic goals with polished, high-quality work.

What you get

• Coding projects: debugged and well-commented Python, SQL, JavaScript or TypeScript code for your assignment or thesis requirements
• Linux support: guidance with advanced shell scripting, system configuration or tools
• Probability & statistics: step-by-step solutions, visualizations and statistical analysis
• Research assistance: help with structuring, coding or interpreting results for thesis projects, along with insights on best practices and clear documentation

Why you might need it

• Improve academic performance: understand key concepts and submit high-quality work
• Save time: get efficient solutions while focusing on learning

What I do

• Review your assignment or project guidelines to understand the requirements
• Provide solutions with comments, explanations and, probably, visual demonstrations

Prerequisites

• Detailed instructions for the assignment, coursework or thesis goals
• Access to any relevant datasets, systems or materials for your project
• Any tools, frameworks or environments required by your institution

Preliminary pricing

$50–$300 for coding tasks, statistical and mixed problems. $500+ for complex research projects. These prices are mostly determined by urgency.

Payment model

Payment is always on an hourly basis.

17

Machine learning models

Using machine learning, I can help you to solve a wide range of problems related to transforming data into useful information by automating decisions and predicting outcomes, giving you a competitive edge. Whether you represent a company or just working on a standalone project, I can implement, update or maintain (ongoing or ad-hoc) statistical algorithms, neural network models or ML pipelines for classification, regression, clustering, dimensionality reduction, anomaly detection and other (mixed) problems.

This service is extremely diverse and will be divided into more specific sub-types in the future. At the moment, all details are subject of discussion.

Examples

👀RS pipeline       

An implementation of a full recommender system pipeline using PyTorch, Elasticsearch, Redis, Flask, Feast and Triton: from offline data processing and model training to online serving and prediction.

👀 RS pipeline    — An implementation of a full recommender system pipeline using PyTorch, Elasticsearch, Redis, Flask, Feast and Triton: from offline data processing and model training to online serving and prediction.

📈Etsy       

Time series forecasting system leveraging ARIMA/SARIMA, Prophet, LSTM networks and ensemble methods along with automated feature engineering (lag features, rolling statistics, date-time components) for real-time predictive analytics on simulated streaming data

📈 Etsy    — Time series forecasting system leveraging ARIMA/SARIMA, Prophet, LSTM networks and ensemble methods along with automated feature engineering (lag features, rolling statistics, date-time components) for real-time predictive analytics on simulated streaming data

Preliminary pricing

Price is negotiable.