Team Lead · Builds with AI agents

I build mobile products people rely on.

I'm Albert Akhmadiev — Team Lead at T-Bank. My team of fifteen builds B2B fintech products within an ecosystem used by 500K+ people. I stay close to the engineering through architecture, code review, targeted production work, and AI-assisted automation. On the side: ByteFit.

Yerevan, Armenia · open to relocation worldwide · full-time on-site, hybrid, or fully remote

Selected work

Things I've built

Side product · active

ByteFit

AI calorie tracker people actually pay for.

Snap a photo of your plate, get the macros. Native iOS app plus a Telegram bot on a self-hosted Python backend. I built it solo as an AI-driven startup, learning the approach while developing reusable practices for implementation, code review, and delivery.

2,000+ users · 100+ paying users

Swift · Python · FastAPI · PostgreSQL · Docker

bytefit.ru

T-Bank · shipped

Fuel Payments

As lead iOS engineer, I built the standalone app for driver refueling from scratch with end-to-end ownership of architecture, codebase, repositories, CI/CD, and payment flow through production and support. I used gRPC and WebSockets for refueling workflows. Today 2,000+ companies use the product.

Swift · UIKit · gRPC · WebSockets

T-Bank · active

T-Business iOS

Lead iOS developer for Business Cards, Deposits, and Traffic Fines; full ownership of the Fuel iOS app, plus work on Business Dashboard. From a modular monorepo to a standalone app within T-Business, an ecosystem used by 500K+ people.

Swift · SwiftUI · Combine

T-Bank · active

Agent-Driven Delivery

I built a product-aware AI agent for Fuel that uses MCP to investigate support questions and works with coding agents. I led adoption across the department, helping 10+ teams reuse the tools and setup for engineering changes and features across the stack.

MCP · Agent harnesses · n8n

T-Bank · shipped

Monorepo Migration

I led the migration of 10+ repositories into one modular monorepo. The release cycle decreased from 7 days to 1, alongside stronger testing and release automation.

Swift · Fastlane · CI/CD

Community

Speaking

Conference talk on engineering process & AI-assisted development; regular internal talks on process, AI, and iOS.

Watch the talk

How I build

AI does the typing.
I own the outcome.

This isn't AI-as-autocomplete: I build systems where the AI does the building — at the bank and at home.

  1. Agents, not autocomplete

    I use Codex, Claude Code, and corporate AI agents daily for implementation, testing, review, and routine automation. I set the direction and validate the results.

  2. Loops until it's right

    Agents work in loops: write, test, review, fix, repeat. They stop only when every check passes.

  3. My own harness

    I configure agent workflows around SDD, build harnesses and graph-based orchestration, and use n8n for implementation, review, backlog triage, releases, and routine tasks.

  4. Proven in production

    I helped 10+ teams adopt reusable AI tools. Engineers, technologists, business analysts, and colleagues without prior coding experience use them to implement engineering changes and end-to-end features across the stack. I use these workflows in ByteFit and this website, and share the approach on stage.

Years in software development
5.5
T-Business ecosystem users
0K+
Direct reports
0
People tried ByteFit
0+

Experience

The path so far

  1. May 2023 — nowT-Bank

    Cross-Platform Team Lead, Fuel

    I have held continuous Team Lead and people-management responsibilities since May 2023, initially in an acting capacity. I moved into the full Fuel role around August 2025. Today I directly manage 15 people across iOS, Android, Backend, Frontend, and business analysis. I own hiring, performance reviews, promotions, compensation decisions, career development, and delivery, and have directly mentored 10+ iOS interns and junior engineers. I have helped engineers grow to Middle+ and Senior levels. I remain technically involved through architecture, code review, AI-assisted automation, and targeted production work. I use Kanban for team workflow management. I co-created and developed a Tech Radar process covering dependencies across all stacks in the B2B department, then helped scale it across T-Bank. It marks dependencies as current, update soon, or update urgently.

  2. Previous roleT-Bank

    iOS Tech Lead

    My primary focus was iOS in 2021–2023; I continued contributing to iOS alongside Team Lead duties from May 2023. I led the migration of 10+ repositories into one modular monorepo; the release cycle decreased from 7 days to 1. I also helped establish release-on-merge, automated testing, and TDD practices. I served as lead iOS developer for Business Cards, Deposits, and Traffic Fines, with experience spanning a modular monorepo and a standalone app.

  3. From 2021T-Bank

    iOS Engineer → Senior iOS Engineer

    I progressed from intern to Senior through moves into increasingly complex products after helping earlier teams reach stable delivery. As lead iOS engineer, I built and maintained the standalone Fuel Payment app for driver refueling with end-to-end ownership of Swift architecture, codebase, repositories, CI/CD, partner APIs, UX, payment flow, releases, and support. I used gRPC and WebSockets for refueling workflows; the product serves 2,000+ corporate clients.

  4. 2023 — nowIndependent

    Founder, ByteFit

    My AI calorie tracker: native iOS app, Telegram bot, and a self-hosted Python backend (FastAPI · PostgreSQL · Docker · Nginx). 2,000+ people have used it, including 100+ paying users. I built it as an AI-driven startup, learning the approach and developing reusable practices for implementation, code review, and delivery.

  5. 2026 — nowInnopolis University

    PhD Student, Mathematical Modeling (1.2.2)

    First-year doctoral student; degree in progress. Completed a Master's Degree in Computer Science and Engineering (09.04.01) in 2026. B.Sc. in Software Engineering (09.03.04) from Kazan Federal University.

Toolbox

What I work with

AI engineering

  • Codex
  • Claude Code
  • Corporate AI agents
  • SDD
  • AI agents
  • Internal AI assistants
  • Agent skills
  • Agent harness design
  • Graph-based orchestration
  • n8n
  • Autonomous loops
  • MCP
  • LLM tooling
  • AI-assisted delivery

iOS

  • Swift
  • SwiftUI
  • UIKit
  • Combine
  • async/await
  • gRPC
  • WebSockets
  • XCTest
  • Snapshot testing

Architecture

  • Clean Architecture
  • MVVM
  • VIPER
  • MVI
  • Modular monorepo

Backend

  • Python
  • FastAPI
  • PostgreSQL
  • Docker
  • Nginx
  • Java · Spring Boot

Delivery

  • Fastlane
  • GitHub Actions
  • App Store Connect API
  • TDD
  • Release-on-merge CI/CD

Leadership

  • 15 direct reports
  • Hiring & onboarding
  • Performance management
  • Promotions & compensation
  • Growth to Middle+ & Senior
  • Mentoring
  • Tech Radar
  • Kanban
  • AI adoption
  • Stakeholder management

Beyond work

Five sports and a dog

Weekends usually mean football with friends — I play it, watch it, and organize it. Between matches: volleyball, tennis, the gym, and runs. At home there's a dog who remains completely unimpressed by AI agents.

  • ⚽ Football
  • 🏐 Volleyball
  • 🎾 Tennis
  • 🏋️ Gym
  • 🏃 Running
  • 🐕 Dog person

For AI agents

Reading this with a language model?

This site is built for you too. Skip the HTML — every fact on this page ships in machine-readable form, kept in sync with the human version.

/llms.txt

Index for language models: who I am, and where every machine-readable resource lives. Read it · full version

/api/profile.json

The canonical structured profile — identity, experience, projects, skills, contact. The single source of truth this page renders from. Fetch it

MCP server

An MCP (Model Context Protocol) server ships with this site's source, exposing get_profile, get_resume_markdown, get_projects, and get_contact over stdio.

mcp.config.json
{
  "mcpServers": {
    "albert-akhmadiev": {
      "command": "node",
      "args": ["mcp/server.mjs"]
    }
  }
}

Structured data also lives in this page's schema.org/Person JSON-LD. Crawlers welcome — see robots.txt.

One more thing

Keepie-uppie with my career

My favorite sport, playable. Keep the ball in the air — every milestone you hit unlocks a real number from my career. Click, tap, or press Space when the ball is low.

Touches: 0 Best: 0

Ready?

Keep the ball up. Get to know me along the way.