FG_/DEV

Portfolio / 2026

Fedor Gofman

Full‑stack
developer.
ERP & AI.

I take products from zero to real production use: manufacturing ERP systems, voice AI assistants, integrations at the protocol level rather than just REST API. 6+ years in development.

Fedor Gofman

Novosibirsk / Remote

01 / About

I specialize in taking a product from zero to real production use — I designed and single-handedly shipped several ERP systems for manufacturing companies: architecture, database, role-based access, warehouse and production accounting. These systems now run in production at clients; they didn't stay pilots.

A separate focus is AI integrations: voice assistants built on Yandex SpeechKit Realtime + Asterisk telephony, dialog-analysis pipelines with LLM scoring. I also have experience with less typical tasks — for example, implementing the TonConnect protocol by hand (decrypting bridge messages via NaCl), which requires working a level below the usual REST API.

I actively use AI coding agents (Claude Code) as my main development tool — I assign tasks, review the diff, require tests. I prefer to understand the business process before writing code.

02 / Experience

02.2026 — present

01

Ivolga

Full-stack Developer (AI integrations, ERP systems)

  • Designed and built 2 independent manufacturing ERP systems (woodworking, steel structures): role-based access (up to 7 roles), a multi-stage production cycle from order to shipping, warehouse accounting
  • In-app docx/xlsx editing integrated via OnlyOffice Document Server
  • Self-hosted infrastructure on Supabase (PostgreSQL + PostgREST + Auth + Storage + Edge Functions), multi-tenancy
  • Both systems shipped and put into production use at clients
  • A line of voice AI assistants (IVR) for three business verticals (tire service, pawnshop, tech support): Yandex SpeechKit Realtime + Asterisk telephony (AudioSocket/ARI), automated appointment booking, reserving stock in the warehouse, CRM (Bitrix24) and mapping-service (Yandex Maps) integration
  • A dialog-analysis system for retail: speech recognition (Whisper / Yandex SpeechKit) → speaker diarization (pyannote.audio) → LLM scoring against the sales script (Ollama) → dialog-path visualization for service-quality control
  • In parallel, contributing to an AI module (an AI-agent panel with multiple LLM providers, chat, persistent history) for a cross-platform desktop app on Rust/Tauri/Vue

07.2024 — 09.2025

02

Independent project

Software Engineer (Automation / Reverse Engineering)

  • Implemented the TonConnect protocol (TON wallet connection) by hand: decrypting bridge messages between the dApp and the wallet via NaCl box decryption, recovering the secret key — reverse engineering of a closed data-transfer protocol
  • Built browser-automation clients on Playwright/Patchright for Telegram Mini Apps: client-environment emulation, network-traffic interception (including the TonConnect bridge), crypto-wallet integration (TON, Solana)
  • Integrated an anti-bot bypass system (captcha solvers: hCaptcha, reCAPTCHA, FunCaptcha, Turnstile, GeeTest)
  • Automation of web-application interaction scenarios: Python, Selenium, Node.js, JS scripting
  • Optimized the system for high load: stable parallel operation of up to 40,000 concurrent sessions/instances

02.2023 — 06.2024

03

Simex (private pension fund)

Software Engineer

  • Built a web system for generating commercial proposals for the sales department (React + Node.js): automatic docx generation, converting amounts to words
  • Developed algorithms for identifying substance composition from spectrometry data (Python): finding the closest reference spectrum

02.2021 — 02.2023

04

Independent projects

Frontend Developer / Automation Systems Developer

  • Built a blockchain dApp (WAX/EOS): React + Redux Toolkit, crypto-wallet auth, game modules (crafting, market, staking) — about a year of active development on a team
  • React frontend development for an external project (~3 months)
  • Built browser-automation systems: Python, Node.js, Selenium — automating web-application interaction scenarios
  • Reverse engineering of frontend client logic, analysis and bypass of protection against automated access (anti-bot systems, behavioral checks)

02.2020 — 02.2021

05

DataKrat-S

Developer

  • Started while at university — first an internship, then employment (~1 year)
  • Built bash scripts for automatic infrastructure deployment and updates (a homegrown CI/CD): a Node.js/Express server received GitHub webhooks and ran the deploy of the NestJS backend and frontend
  • Sole developer of a restaurant loyalty system, from scratch: database design (MongoDB), microservice architecture on NestJS (auth, menu, cart, loyalty cards, API gateway), React Native frontend
  • Integration with the r_keeper POS system: pulling the menu via the POS's native XML HTTP API, plus a custom Delphi module linked directly into the POS to accept loyalty cards
  • CRM integration (calls/deals over Asterisk telephony), parsing PDF invoices, customer face recognition
  • Delivered the full feature set: pre-order, menu loaded from the POS, loyalty program

03 / Stack

TypeScriptNode.jsNestJSFastifyNext.jsReactPythonFastAPIPostgreSQLMongoDBDrizzle ORMSupabaseDockerNginxAsteriskYandex SpeechKitLLM integrationsTelegram Bot APISolidity

Non-trivial engineering cases — on a dedicated page

View projects →