Nikolay Kokov
Full-stack Developer (Python, React) AI / LLM integrations Project / Product Manager
I build web products end to end and run the projects around them. Full-stack developer since 2016 (Python, FastAPI, Django, React), leading projects and teams since 2021. I either launch my own products or work directly with clients: gather requirements, write specs, plan the work, set tasks for the team and ship to production.
- since 2016in software development
- PythonFastAPI, Django + React
- 5–7× fasterdata collection
- 20× morematches with LLM
Full-stack Developer · Python, React
About
Full-stack developer since 2016. Core stack: Python (FastAPI, Django) and React. Recent work: data collection services, scraping bot-protected sites, LLM integrations.
I build projects end to end: from requirements and architecture to production deployment and support. I talk to clients directly and can lead a team.
English C1.
Strengths
- Ship a service to production solo: scraper, API, frontend, deployment
- Reliable data collection where sites block automation
- Use LLMs where they give a measurable result
- Understand the business goal and work with clients directly
Skills & tools
- Backend
- Python, FastAPI, Django, Celery, asyncio, PHP (Yii2)
- Data & scraping
- Playwright, Camoufox, residential proxies, ADB
- Databases
- PostgreSQL, MySQL, Redis
- Frontend
- React, Vue, JavaScript, TypeScript, WebSocket
- Mobile
- Flutter/Dart, Firebase
- AI
- OpenAI, Anthropic, OpenRouter APIs, ComfyUI, RunPod
- DevOps
- Docker, Kubernetes, Grafana, GitHub Actions, Codemagic
Experience
-
Finds stores on Google Shopping that sell the client's products, collects their prices and shows where the client stands on the market. First client: a European consumer electronics and home appliances retailer, 12,000+ products.
- built the service solo, from scraper to production deployment
- browser-based collection (Playwright) with a residential proxy pool and profile rotation, stable despite Google's anti-automation
- parallelized collection across multiple browsers: 5–7× faster per product
- added LLM-generated search queries: 20× more matches for catalogs without specs
- matching competitor products by barcode (EAN) and attributes
Python, FastAPI, Celery, PostgreSQL, Redis, Playwright, React, Docker, GitHub Actions
-
- designed and built the service, from scrapers to production deployment
- wrote collectors for 4 source types: RSS, Telegram (via an Android device over ADB), Discord, Substack (headless Playwright/Camoufox)
- connected LLMs (OpenRouter): daily per-channel digests, a cross-source overview and automatic infographics
- built a Django admin panel for managing channels and sources
Python, Django, Playwright, Camoufox, ADB, OpenRouter
-
- developed and refactored the platform backend and API (FastAPI)
- built an API for managing servers and test nodes for crypto projects, covered with unit and integration tests
- took a Telegram bot for photo and video generation from a vibe-coded prototype to production
- deployed ComfyUI on RunPod (On-Demand and Serverless) and connected it to the bot API and Telegram WebApp; bot infrastructure on Kubernetes, logging in Grafana
- built an MVP of an AI avatar chat platform; payment integrations, AI agents, prompt engineering
Python, FastAPI, asyncio, Redis, Docker, Kubernetes, Grafana, ComfyUI, RunPod
-
Matches supplier price lists to tender bids: processing time per bid cut from a week to one day.
- partly developed the backend; managed the project and client communication
- vector search matching bid items (up to 10,000) against several supplier price lists, picking the best price
- different LLMs per step: lightweight ones to shortlist by name, stronger ones to match by specification; irrelevant fields filtered out
- match review form (confirm, reject, correct manually) and automatic commercial proposals
- scraping product specs from supplier websites (VseInstrumenty.ru and others)
-
- built the frontend of the platform and the CRM for operators, boosters and admins
- designed the system architecture and functionality with the backend developer, based on the client's requirements
- built a product builder with conditional option logic
- implemented real-time UI updates (WebSocket) and a support chat in the admin panel
- integrated Stripe and PayPal
- wrote specs, coordinated the designer and backend developer, set up database backups
PHP (Yii2), React, MySQL, WebSocket
-
- built the frontend and a CMS for the catalog and content
- implemented SEO and Core Web Vitals optimization and new page layouts
PHP (Yii2), React, Docker
Other projects
- AI companion startup (NDA), Product Manager2025 – present
Launched the product from zero: economics, roadmap, team. Built the MVP on LLM APIs (OpenAI, Anthropic, OpenRouter, Grok); integrated payments: WATA (SBP), Capitalist and StreamPay (crypto), Telegram Stars.
- Pesok beach sports centerMay – Sep 2026
Website redesign and upgrade of all internal packages: frontend migrated from Vue 2 to Vue 3, Django backend, API between them.
- brandrecon.ai2025
Live AI service measuring brand mentions across the web and search. Built the React frontend: page components and admin panel dashboards (React, FastAPI).
- “Save Humanity” audio play app2023–2024
Mobile app: architecture, development, App Store and Google Play release, maintenance (Flutter, Firebase, Codemagic).
- Astreya and Origami mobile apps2021–2022
Dental clinic and food delivery apps: spec, UI, store release. 1C integration, YooKassa payments and notification admin for Astreya (Flutter, Django, Firebase).
- Independent Reserve (crypto exchange, Australia)2022–2023
New website version: custom Gutenberg blocks and PHP templates (WordPress).
- Game portals for Polyus2021, 2023
Browser games with leaderboards and admin panels. Developed and led teams of 2–3 developers (Django, Yii2, JS).