Michael Batrakov

ML DevOps Test Software Data

# Summary

I’m a full-stack engineer and AI infrastructure specialist with experience in building scalable systems, automating workflows, and creating data-driven solutions. My work includes developing self-service platforms, AI agents, and optimizing processes in both software and hardware. I focus on practical, efficient solutions and adapt quickly to new technologies.

# Skills

Category Skills
Programming Python, C/C++, MATLAB, TypeScript, JavaScript, R, Bash/Shell/CShell Scripting
Frameworks & Libraries Pandas, NumPy, Matplotlib, Seaborn, Django, React JS, Next JS, Svelte, Flask, Regex, SQLite, FastMCP
Databases PostgreSQL, MySQL, Firebase, MongoDB, SQLite
Machine Learning TensorFlow, PyTorch, Scikit-learn, Keras, Fundamental and abstract understanding of CNNs, RNNs, hyperparameter tuning, regularization, optimization techniques, and best practices for structuring and scaling machine learning projects.
DevOps & Tools Git, Docker, CI/CD Pipelines, Linux Server Management, Gunicorn, Nginx, Apache, WebSockets, Agile, Jira, Confluence, LabVIEW, ROS, Modbus, Altair Monitor
AI Infrastructure RAG (Retrieval-Augmented Generation), MCP (Model Context Protocol), API Security, Prompt Engineering
Web Development Full-Stack, RESTful APIs, Authentication Systems (Okta, LDAP, OAuth), Caching, Optimization, Software Architecture
EDA License Management License Server Administration, Usage Auditing, Automated Alerting, AI-Driven Support Systems, FlexLM
Certifications & Courses Neural Networks & Deep Learning, Statistical Consulting, Deep Learning Specialization (Stanford Online)

# Experience

Download Resume

# Projects

Neural Network Visualizer



Project Link


  1. Developed a live neural network visualization tool using TypeScript, allowing real-time interaction with neuron activations through mouse movement and drawing.
  2. Implemented forward propagation from scratch, demonstrating a strong understanding of neural network mechanics.
  3. Integrated weights trained in Google Colab into a custom backend function, enabling dynamic network building and visualization in the frontend.
  4. Enhanced user experience by providing intuitive visual feedback of neuron activations and network behavior.
  5. 2D graph label predictor and hand-written digit recognizer.
  6. Technologies: Typescript, NextJS, Python, Numpy, Tensorflow, Pandas, Matplotlib

3D Portfolio


Project Link


Portfolio created in an early morning foggy redwood forest.

Simple Todo List


Project Link


  1. Simple yet powerful todolist that utilizes the browser's local database to store tasks (Indexed DB).
  2. One input can control due date, priority, cycle, and label of tasks.
  3. Tasks that are finished are automatically archived, unless it repeats.

Portfolio (this)


  1. This portfolio site is running on a raspberry pi.
  2. Technologies: Python, Nginx, Gunicorn, Flask, and Postgres.