Daniil Maklakov

Backend developer — Python/FastAPI · TypeScript/Fastify · CRM & API integrations

Daniil Maklakov

I build the plumbing between products: API gateways, OAuth2 flows, webhook pipelines and CRM synchronisation that run in production for real clients. Freelance backend developer since late 2025.

Originally a physics engineer, I moved into software through a computer science degree in Strasbourg. I work remotely, deploy and maintain what I ship, and use LLMs daily — both as tools and inside products. Based in France. Open to backend and AI-integration roles.

Experience

Backend Developer — Optima Fence and Deck Feb 2026 – present

Freelance, remote. CRM integrations in Python/FastAPI, deployed and operated on AWS.

Backend Developer — H Alliance May – Jun 2026

Freelance, remote. Meta Business and Instagram APIs synchronised with HubSpot CRM in TypeScript.

Backend Developer — Thumbtack API gateway Nov 2025 – early 2026

Freelance, remote. OAuth 2.0 gateway and synchronisation pipeline across Thumbtack, HubSpot and Zoho. Still in production.

BSc Computer Science — University of Strasbourg 2023 – 2026

Preceded by a BSc in Engineering Physics, Kazan Federal University (2016 – 2020).

Projects

Replaced the client's no-code Zapier automations with Python pollers and webhook handlers, removing the dependency without changing anything on the CRM side. Keeps GoHighLevel and ServiceTitan in two-way sync: 13+ webhook event types, status mapping, deduplication, plus lead intake from Yelp, Thumbtack and HouseCall Pro. Sanitized production codebase.

halliance-platform TypeScript

Integration platform monorepo: Fastify gateway with HMAC-preserving request forwarding and typed API clients for HubSpot, Aircall, eBay, Meta and Zoho. Sanitized production codebase.

Production OAuth 2.0 reverse proxy for the Thumbtack API: full token lifecycle, FastAPI and PostgreSQL, Dockerised behind nginx, with an internal admin panel. Built in late 2025 and still running unattended.

drawparty TypeScript

Real-time multiplayer drawing game with an AI Judge mode: a vision LLM watches the canvas and guesses drawings live. Next.js, Electron, socket.io, Prisma. Built in a 9-person team.

Decision tree implemented from scratch in NumPy, with per-instance explainability via Monte Carlo Shapley-value attribute contributions.

Stack

Backend: Python (FastAPI, httpx), TypeScript (Fastify, Node.js), PostgreSQL

Integrations: OAuth2, REST APIs, webhooks, CRM sync, LLM APIs

Ops: Docker, Linux, nginx, Git, CI/CD

Contact

daniil.maklakov8@gmail.com

github.com/McDonnies · linkedin.com/in/maklakov-daniil

Download CV (PDF)

English · French · Russian