Yuriy Chukaev — Senior Software Engineer (AI-native)

Track record

Proof, in numbers

10+ yrs

commercial software engineering

69,000+

sales calls reviewed by an AI pipeline I built — every conversation, not a spot-check

43,000+

leads routed to the right person in seconds — none left going cold in a queue

40,000+

parcels/day at peak on a logistics platform I built and ran for 3+ years

300+

companies integrated over a product's lifetime

14,000+

SKUs selling across 6 storefronts in two countries on a platform I built

I own delivery end to end — architecture, code, third-party integrations, deployment and long-term evolution. You talk to the engineer who writes the code. And shipping to spec is table stakes: what I optimize for is the outcome the software exists to produce — revenue kept, hours returned, decisions made on real numbers — not an admin panel that demos well.

See the full track record — ~25 projects over 10 years →

Case 01 · On-demand operations + AI intake

From paper notebooks to a live operations app

Problem

An on-demand florist-retail business ran every order through paper notebooks and chats. Managers matched florists and couriers by hand, order by order — the bottleneck any on-demand service hits as it grows. Orders slipped between apps, and after-hours customers went unanswered — free to buy from whoever replied first.

Approach

A role-based PWA (installs like a native app) with a separate cabinet per role — florists, couriers, managers, accounting — covering the whole order lifecycle: intake → "take the job" → assembly → photo confirmation → automatic payment request. On top, an AI consultant answers customers across three channels, checks live stock and price, upsells, and opens the order automatically; the owner edits how it talks right in the app, every change versioned.

Stack: React + Vite + TypeScript · FastAPI · async SQLAlchemy · PostgreSQL · CRM API (amoCRM v4) · LLM consultant · PWA

Impact: no order lost between chats, customers answered the moment they write — 24/7 — and the whole order flow visible in one system; the owner, not the vendor, controls how the assistant talks.

Read the case — architecture & decisions →

Case 02 · E-commerce & integrations at scale

A pharmacy chain, fully online

Problem

A regional pharmacy chain (13 stores across two countries, 6 legal entities) needed to sell online — with live stock per store, regulated products, local pricing and lawful fiscal receipts. Until that worked, online revenue was off the table — and any shortcut risked unlawful receipts and oversold regulated products.

Approach

I migrated the chain off a boxed CMS onto a custom PHP 8.3 engine with deep two-way integration to two pharmacy ERPs, payment routing across legal entities, fiscalization, delivery integration, messenger bots, and a multi-region plus second-country launch.

Stack: PHP 8.3 · MySQL · REST & webhook ERP integrations · payment gateways · fiscalization · delivery API

Impact: the chain sells online in two countries — 14,000+ products across 6 regional storefronts, stock honest per store, every payment landing in the correct legal entity with a lawful receipt; the second-country launch was a configuration step, not a rebuild.

Read the case — architecture & decisions →

Pharmacy e-commerce catalog, 14,000+ products

Case 03 · Production AI & automation at scale

AI that listens, scores and fills the CRM

69,000+

calls analyzed in production

A production AI product (my own) listens to calls and chats, fills the CRM automatically, scores quality and flags risks — so the owner gets every conversation reviewed, not the 5–10 a day a human controller manages. In its first 2.5 months it analyzed 69,000+ calls (1,200+ hours of audio) and ran 31,000+ automated checks, isolated per client.

Stack: Python · speech-to-text · LLM scoring (Anthropic / Gemini) · CRM autofill

43,000+

requests routed in production

A distribution engine that routes each inbound request to the right available person in seconds, by configurable rules — load, skills, schedule, priority — so no lead sits unclaimed going cold. In production since 2022, with a client I have supported since 2018.

Stack: Python · rules engine · CRM & telephony integrations · webhooks

Read the case — architecture & decisions →

These systems run inside clients' CRMs, so I show them anonymized here and live on a screen-share call. Other work in my track record: an amoCRM widget platform (23 sub-widgets on a custom PHP backend) and a booking PWA with end-to-end marketing attribution for a medical clinic. I also build my own tooling — including a browser-automation layer that lets an AI agent drive a real browser for everyday workflows.

Why now

Why I'm opening to a role after a decade on my own

For about ten years I've run my own studio, owning delivery for many clients at once. I'm not leaving anything on fire — the business runs. I'm choosing depth over breadth: to go deep on one product, with one strong senior team, and build something that compounds over years instead of switching context every hour.

Based in Brazil with a young family, I value a stable footing to do my best work from, and I want to work alongside people as strong or stronger and learn from them again. I'd join as a contractor, which keeps things clean on both sides — and a decade of running everything myself is exactly the autonomy a senior remote hire needs on day one.

Stack · how I work · contact

What I bring, and how we'd work

Backend

PHP 8 (custom engines, OOP, PDO), Python (FastAPI, Flask), Node.js

Frontend

JavaScript, TypeScript, React, PWA (service workers, web-push)

Integrations & APIs

REST APIs, webhooks, CRM Widget SDK & API (Kommo / amoCRM), payment gateways, ERP & delivery APIs, telephony (SIP / VoIP), IMAP / SMTP

AI & LLM

Production LLM pipelines: structured outputs, speech-to-text, context retrieval from CRM/DB, prompt engineering with reviewed-example regression; Anthropic Claude & Google Gemini APIs (also OpenAI). Multi-agent orchestration and browser tool-use in my own internal ops platform. AI-first delivery — I build with LLMs daily (Claude Code, Cursor)

Data & Infra

PostgreSQL, MySQL, Redis, Docker, Nginx, Linux, CI/CD

You talk to the engineer who writes the code — no juniors, no hand-offs. I take a project from an ambiguous spec to production and keep it running and evolving. Weekly demos, async-first (written updates, PRs), comfortable on calls.

Availability & engagement

Available now — remote contract (Brazil PJ / independent contractor), invoice in USD via Wise / Payoneer. Worldwide, US-hours friendly. No visa sponsorship or EOR needed.
Languages: English — professional working (CEFR B1+) · Portuguese — basic · Russian — native.

mr.chukaev@gmail.com WhatsApp +55 13 99620 3533 LinkedIn Telegram @ychukaev CV (PDF)