38,000+
employees
worldwide remote
AI Product Engineer
LLM features that survive production — retrieval, tool calling, structured outputs — shipped end to end on React and FastAPI
LLM Assistant
Internal AI Product
Image Gen UI
Design Workflow Tool
Angular to React
Migration Profile
Focus
Specialised in the part of AI products that breaks
Most teams can get an LLM demo working. I build the layer underneath: retrieval that returns the right document, tool calls that fail safely, and an interface people open on a Monday morning.
01
AI Reliability
Structured outputs with schema validation, tool calling that handles partial failure, async orchestration and streaming that survives reconnects
02
Full Product Surface
I ship the React interface and the FastAPI service behind it, so AI features do not stall on a backend handoff
03
Legacy Frontend Migration
Angular and Svelte to React in production, without freezing feature delivery
Experience
Recent roles and progression
Official dates come from current employment records. Public copy keeps the focus on shipped work, ownership, and NDA-safe scope
T-Bank
Frontend Engineer, Copilots
Nov 2025 - Present
Shipping features on an internal LLM assistant used by 38,000+ employees, across a mixed React and Svelte frontend and the FastAPI services behind it, owning delivery from the interface through to the backend rather than stopping at a handoff
Shipped product work in a mixed Svelte and React codebase for a widely adopted internal AI assistant, helping move new product logic toward React-owned areas
Continued delivering internal AI tooling features end to end, including UI behavior, backend fixes, test coverage, and code quality guardrails
T-Bank
Frontend Engineer, AI Image Generation
Jun 2024 - Oct 2025
Grew from junior ownership into broader product responsibility in internal AI and experimentation teams, with Angular as the main frontend surface and FastAPI as the usual backend support layer
Owned substantial parts of an internal image-generation UI for designers and its adjacent backend workflows, including generation flows, boards, reactions, and async status handling
Regularly crossed the frontend/backend boundary when features required FastAPI APIs, background consumers, persistence, or service integrations
T-Bank
Software Engineering Intern
Jan 2024 - Jun 2024
Started in an AI-focused internal product environment and quickly moved from internship tasks to production feature ownership
Joined the internal AI product stream and began shipping production changes in Angular and FastAPI
Built the foundation for later fullstack ownership by working directly with product flows instead of isolated training tasks
Qoollo
Software Engineering Intern
Sep 2023 - Jan 2024
Worked on a web app that compared packaging imagery to detect counterfeit products and prepared Python learning materials for School 21
Continued a university practice project into an internal internship context, refining image-comparison screens, admin flows, and uploaded-image editing
Prepared Python study materials and practical exercises for educational programs alongside product work
Case Studies
Internal product work, shown through ownership and systems thinking
The code for these products is not public. What I can show is the product surface, architecture, and execution scope behind them
Internal design workflow tool
1,000+ internal users
Image Generation Platform UI
Internal web application for an AI image-generation platform used by designers who needed a practical interface for day-to-day creative work. The product provided board-based organization, editing workflows, and async task management on top of a company-wide generation API
My role
Primary ownership on the Angular frontend, with selected end-to-end delivery across the FastAPI backend when features required server-side work
Angular
TypeScript
NgRx
RxJS
Taiga UI
OAuth / SSO
Built and evolved a board-based workspace where designers organized generations into project spaces instead of a flat task list
Implemented drag-and-drop between boards, image reactions, and editing flows for multiple generation modes
Designed a priority-queue polling model on the frontend to schedule status checks for asynchronous generation tasks without aggressive polling
FastAPI backend-for-frontend
Image Generation BFF
FastAPI service that connected the frontend to the company-wide image-generation API, handling authentication, task lifecycle data, and downstream persistence for the interface
My role
Designed and implemented the original BFF surface, then kept extending it as product features moved beyond frontend-only scope
FastAPI
Python
PostgreSQL
SQLAlchemy
Kafka
S3
OAuth / SSO
Alembic
Integrated corporate SSO for end users while routing generation requests through a service-level backend path
Built an event consumer that processed completed tasks and coordinated persistence across the database and object storage
Added backend support for user-facing features such as reactions, board operations, and image post-processing flows
Internal LLM assistant platform
38,000+ employees
Internal AI Assistant
Company-wide internal LLM assistant used by 38,000+ employees. The product is built on a mixed Svelte and React frontend with a FastAPI backend and is progressively moving toward a React-first architecture without pausing feature delivery
My role
Contributed across frontend architecture, quality tooling, and feature delivery while the product moved from a legacy frontend toward React
React
TypeScript
Svelte
FastAPI
Vitest
Biome
Oxlint
Playwright
Worked in a mixed frontend stack, helping move product logic into React-owned areas while maintaining the existing Svelte surface
Shipped product features and supporting backend fixes in parallel with migration work instead of treating them as separate tracks
Helped establish stricter frontend quality rails around new React code while the product still carried legacy surface area
Public Work
Selected repos that show how I build
Small, finished pieces of tooling and product work that are open enough to read end to end
GitHub Repo
StudHack
Platform for hackathon teams, events, and career communities with user profiles, team workflows, notifications, and Dockerized UI/API delivery
Angular 21
TypeScript
ASP.NET Core 8
PostgreSQL
GitHub Repo
FastAPI MVP
A reusable FastAPI bootstrap package with Mongo and S3 helpers, metrics, and stricter environment management for new service setup
FastAPI
Python
MongoDB
S3
Prometheus
GitHub Repo
OpenAI Async Functions
Async Python library for OpenAI-compatible tool calling, with Pydantic request and response models, runtime tool registration, and structured output helpers
Python
AsyncIO
Pydantic
Tool Calling
OpenAI-compatible APIs
GitHub Repo
lint-gost-tex
CLI linter for LaTeX documents against GOST-style formatting rules, with checks for references, lists, captions, forbidden abbreviations, Unicode, and RU/EN spelling
Python
CLI tooling
LaTeX
Text linting
Supporting Project
AIDoc
Supporting AI documentation-generation project combining a .NET CLI and a FastAPI backend
Stack
Comfortable across product, platform, and delivery
Frontend
TypeScript
React
Angular
Svelte
NgRx
RxJS
Tailwind CSS
Taiga UI
Svelte
Tailwind CSS
Backend
Python
FastAPI
SQLAlchemy
PostgreSQL
Kafka
S3
Pydantic
Alembic
Python
SQLAlchemy
PostgreSQL
Kafka
Product and Platform
OAuth / OIDC
Docker
Kubernetes
Prometheus
Grafana
Sentry
CI pipelines
Playwright
Docker
Kubernetes
Prometheus
Grafana
Sentry
Playwright
Working Modes
End-to-end feature delivery
Internal AI tools
Migration programs
API integrations
Design-system alignment
Cross-functional collaboration
Education
Formal education and long-term signals
Bauman Moscow State Technical University
Bachelor’s in Software Engineering (IU7)
Samsung IT School
Android Development in Java, diploma with honors
Working languages
- Russian — Native
- English — IELTS Academic 7.5 (C1)
Highlights
Signals beyond the day job
2023
Finopolis 365 AI Track
Prize winner in a youth AI-focused hackathon stream
2023
AI Talent Hackathon
Built product prototypes in an AI-generated content hackathon environment
Contact
If you are hiring, start here
I’m most useful on teams shipping AI features that have to work for real users: retrieval, tool calling, and the async plumbing underneath, with the React interface and the FastAPI service behind it owned by the same person
