Senior Full-Stack TypeScript Engineer & Tech Lead

Vladimir Fedorov

I build AI products — LLM pipelines, AI agents and computer vision — on top of 19 years of production web development with TypeScript, React, Next.js, Node.js, PostgreSQL and AWS.

Co-founder and CTO of ShelfSignal, where I designed the architecture and built the pipeline that measures how ChatGPT, Claude, Perplexity, Gemini and Grok recommend brands and which sellers their shopping agents choose. Before that: Route4Me, a 10M-daily-users data platform at Bright Data, and twelve years at TypingMaster leading Typetastic, used by about 4M students. As a freelancer: 75 projects for 62 clients with a 5.0 average rating, including MVPs for Siemens, Seagate, L’Oréal and Citroën.

Core expertise

AI products on a full-stack foundation

AI / LLM

OpenAI, Anthropic and Gemini APIs, multi-LLM evaluation pipelines, AI shopping agents, structured outputs, prompt engineering, AI-assisted development (Claude Code)

Computer vision & ML

TensorFlow.js, MediaPipe, OpenCV.js, Deepgram speech recognition, PySpark, Spark MLlib, DuckDB

Frontend

TypeScript, React, Next.js, Redux, Vue.js, Angular, Tailwind, SCSS, Material UI

Backend

Node.js, Express, NestJS, Python, PHP (Laravel, Symfony), REST APIs, GraphQL, Socket.io

Databases, cloud & DevOps

PostgreSQL, Supabase, MongoDB, MySQL, MariaDB, Redis, Firebase, AWS (ECS, CloudWatch, SNS), Docker, CI/CD, Playwright, ElasticSearch

Leadership

Team lead for frontend, backend, QA and DevOps engineers, architecture, code review, mentoring, Agile/Scrum

Selected work

AI products and platforms used by millions

Recent AI work first, then selected projects from employment and the WebCraft portfolio.

ShelfSignal home page: the Order Shelf report for the magnesium categoryshelfsignal.pro

ShelfSignal

Co-founder & CTO. Multi-LLM pipeline that measures how ChatGPT, Claude, Perplexity, Gemini and Grok recommend brands, plus an AI shopping-agent simulation that records which seller gets the order.

CTO · TypeScript · Next.js · Node.js · PostgreSQL · OpenAI, Anthropic & Gemini APIs
AI photobooth kiosk at a shopping-mall event, guest scanning the QR code to download the portraitLive events · Singapore

AI Photobooth

Generative-AI portrait kiosk for corporate events: QR check-in, branded AI portraits with outfit and car selection, animated takeaways, automatic retry and download by QR. About 80–100 guests a day.

Next.js · PostgreSQL · Replicate · WAN 2.2 · TensorFlow.js · ffmpeg
Cultural Quest admin panel with the points map and a phone screen running pose recognitionAR quests · Seoul, Dubai, Berkeley pilots

Cultural Quest

AR quests in the mobile browser with on-device AI: object and painting recognition, pose matching and voice tasks. Video never leaves the phone.

OpenCV.js · MediaPipe · TensorFlow.js · Deepgram · Next.js · PWA
SparkURL pipeline run report: stages, model metrics and URL class distributionCarnegie Mellon University

SparkURL — ML pipeline for CMU research

PySpark, Spark MLlib and DuckDB classify billions of clickstream URLs into purchases and ad clicks on an HPC cluster, with a validation report on over- and under-counting.

PySpark · Spark MLlib · DuckDB · Parquet · SLURM · Python
Bright Data website screenshotbrightdata.com

Bright Data

High-load data search and storage platform with about 10M daily unique users. Redis caching of metrics cut database load by about 30%.

Full Stack · Node.js · TypeScript · React · Next.js · MongoDB · Redis
Route4Me route optimization platform screenshotroute4me.com

Route4Me

Full-stack development of the web portal for a last-mile route optimization platform. Made the legacy portal about 2.5× faster by batching Google Maps requests; built a Geotab Add-In and whitelabel portals for FleetComplete and PowerFleet.

Full Stack · Vue.js · React · PHP/Laravel · MySQL · Google Maps API
Citroën Ami ecommerce checkout screenshotstore.citroen.co.uk/ami

Citroën Ami

Complex EU-market ecommerce application for browsing, configuring, financing and ordering electric vehicles.

React · Next.js · Node.js · Redis · MariaDB
Typetastic product screenshottypetastic.com

Typetastic

Typing platform for schools in the US and Europe: about 4M users, 700+ courses, shipped on web, Windows, Android and iOS. Led a team of 4 developers and QA for twelve years.

Team Lead · Vue · PHP/Laravel · MySQL · Socket.io
EarnApp website screenshotearnapp.com

EarnApp

Portal for earning money online with full-stack product development.

Node.js · React · MongoDB · Redis
Whitehall Training website screenshotwhitehalltraining.com

Whitehall Training

Online healthcare startup with modern frontend and backend architecture.

React · Next.js · Node.js · MongoDB
Snapodds product screenshotsnapodds.com

Snapscreen

React/Next.js frontend for a real-time TV content recognition platform built on AI algorithms, with video stream API integration.

React · Redux · TypeScript · Next.js · Video APIs
Sword Online game screenshotswordonline.io

Sword Online

Browser-based multiplayer game platform with real-time interactions.

HTML5 · JavaScript · Node.js · Socket.io
Mahatat streaming service screenshotmahatat.com

Mahatat

Video streaming service with subscription flows, payments and analytics.

React · MongoDB · Next.js · Node.js
Typerush game screenshottyperush.com

Typerush

Typing development game focused on fast, engaging web performance.

React · Redux · Bootstrap
RawShorts editor screenshotrawshorts.com

RawShorts

Animated video builder for business marketing with editor tools, overlays and dynamic font loading.

HTML5 · JavaScript · Frontend performance
TypingTest product screenshottypingtest.com

TypingTest

Typing test and learning platform with modern UI, analytics and scalable web delivery.

JavaScript · jQuery · Bootstrap · SCSS
TypingMaster product screenshottypingmaster.com

TypingMaster

Educational typing software and web platform with HTML5 applications and backend services.

HTML5 · Node.js · CreateJS · PHP · Java

Experience

Where the work creates impact

01

AI products: LLM pipelines, agents & computer vision

Multi-LLM evaluation pipelines with structured outputs, AI shopping-agent simulations, generative imaging for live events and on-device computer vision in the mobile browser.

ShelfSignal · AI Photobooth · Cultural Quest
02

High-load SaaS & data platforms

Backend performance, caching, API design and database scaling for products with large traffic and complex data flows, from a 10M-daily-users platform to a billions-of-URLs ML pipeline.

Bright Data · Route4Me · Citroën Ami · SparkURL
03

EdTech products at scale

Twelve years at TypingMaster leading Typetastic: course systems, school dashboards, student progress tracking and about 4M users across web, Windows, Android and iOS.

Typetastic · TypingMaster · TypingTest
04

Technical leadership & delivery

CTO and team lead for frontend, backend, QA and DevOps engineers: architecture, code review, mentoring and full-cycle delivery from MVPs to long-running production systems.

75 projects · 62 clients · 19 years

Services

Senior full-stack development for product teams

I help startups and product companies design, build and scale AI products and web platforms from MVP to production.

AI product development

LLM pipelines with OpenAI, Anthropic and Gemini APIs, structured outputs, multi-model evaluation, AI agents and prompt engineering, from prototype to production.

Computer vision & ML applications

On-device vision in the browser with TensorFlow.js, MediaPipe and OpenCV.js, speech with Deepgram, and data pipelines with PySpark, Spark MLlib and DuckDB.

SaaS platform development

Full-cycle development for SaaS products, dashboards, admin panels, subscriptions, analytics and integrations.

High-load web applications

Architecture, caching, APIs and database optimization for products with large traffic and complex data flows.

TypeScript, React & Node.js development

Modern frontend and backend with TypeScript, React, Next.js, Node.js, NestJS, PostgreSQL, REST APIs and GraphQL.

Technical leadership

Fractional CTO or team lead: architecture decisions, code reviews, mentoring, hiring support and delivery planning.

Technologies

Modern web stack for scalable products

TypeScript React Next.js Node.js NestJS Express PostgreSQL Supabase Redis MongoDB MySQL AWS Docker OpenAI API Anthropic API Gemini API TensorFlow.js MediaPipe OpenCV.js Deepgram PySpark Spark MLlib DuckDB Vue.js Angular Tailwind PHP Laravel Python GraphQL Socket.io Playwright ElasticSearch

Testimonials

What clients say

“Vladimir is technically very competent and a great communicator. Super reliable and a pleasure to work with.”

Eric Morhenn, Telly Inc

“Always a pleasure to work with Vladimir; professional, highly communicative and always on top of things.”

Anniina Tammosti, TypingMaster

“Excellent technical skills, fast, efficient and very nice to work with. Can only recommend!”

FAQ

Working with Vladimir

What kind of projects do you work on?

AI products (LLM pipelines, AI agents, computer vision), SaaS platforms, high-load web applications and EdTech products. Most recent work is in TypeScript with React, Next.js, Node.js and PostgreSQL.

Do you build AI products end to end?

Yes. At ShelfSignal I designed the architecture and built the multi-LLM measurement pipeline, the shopping-agent simulation, the rules engine and the client dashboards. Freelance AI work includes generative imaging for live events, on-device computer vision in the mobile browser and an ML data pipeline for Carnegie Mellon research.

What technologies do you use?

TypeScript, React, Next.js, Node.js, PostgreSQL and AWS as the core stack; OpenAI, Anthropic and Gemini APIs for LLM work; TensorFlow.js, MediaPipe, OpenCV.js and Deepgram for on-device vision and speech; PySpark, Spark MLlib and DuckDB for data pipelines.

Do you have team leadership experience?

Yes. I am Co-founder and CTO of ShelfSignal and led a team of 4 developers and QA at TypingMaster for twelve years, covering architecture, code review, mentoring and delivery planning.

Contact

Let’s build something useful and lasting.

I work remotely from Tbilisi, Georgia (GMT+4) and am open to senior full-stack and AI engineering roles, consulting and fractional CTO work.

Email: vladimir@vfedorov.com · Phone / WhatsApp: +995 557 70 92 71 · Tbilisi, Georgia (GMT+4)