Portfolio

Suryateja Kommuri

AI Engineer / AI Solutions Engineer — turning messy workflows into shipped AI products.

Open to AI Engineer & full-stack roles• UK / Remote

I'm an AI engineer who turns ambiguous ideas into working, deployed products. Lately I've focused on retrieval-augmented generation, workflow automation, and applied-AI tooling — systems that answer from real documents with citations, score and prioritise leads from public data, and connect APIs into usable business workflows. I came up through full-stack freelance delivery, so I'm comfortable owning a build end to end, from discovery and prototype to a live URL you can click.

Location: London, United Kingdom • Remote-friendly
Availability: Open to AI Engineer, AI Solutions Engineer, and Forward-Deployed AI roles — UK or remote.

Shipped multiple live AI products in 2025–26, including a geospatial Voice-AI prospecting dashboard and a modular RAG chatbot with source-cited answers.
Strong fit for practical AI teams: can discover the workflow, design the AI layer, build the product, and explain the tradeoffs to users.
Founding / sole-engineer experience: scoped, architected, built, demoed, and deployed end-to-end across frontend, backend, retrieval, automation, and cloud.
Full-stack foundation from 10+ freelance web builds — React, Next.js, Firebase, Supabase, Docker, APIs, and CI/CD.

Email me View résumé

LinkedInGitHub

About

My Journey

I learned to build by shipping real things for real users. Freelancing for small businesses, I delivered 10+ full-stack web apps and sites — owning everything from requirements and design to authentication, data, deployment, and the client feedback loop.

That ownership pulled me toward AI. As a founding/sole engineer I’ve built retrieval-augmented generation systems that answer from real documents with citations, a geospatial tool that scores and prioritises sales leads from public data, and developer automation tooling — taking each from a scrappy prototype to a deployed, clickable product.

I like the messy 0-to-1 stretch where a vague idea becomes a working system. I've also written and tested Ethereum smart contracts in Solidity. Outside of building, I train Thai boxing to keep my head clear and keep an eye on MCP, AI agents, and workflow automation.

0→1

Built solo, end-to-end

10+

Full-stack apps shipped

AI products built

100

Lighthouse PWA score

Role fit

Built for practical AI teams

The strongest match is a team that needs someone who can understand the workflow, build the AI layer, wire the product together, and demo the result without a long handoff chain.

AI Engineer

Built RAG and applied-AI systems with retrieval, citations, OCR paths, evaluation-minded fallbacks, and live demos.

AI Solutions Engineer

Turns messy business workflows into working AI tools: discovery, data model, prototype, demo, and deployment.

Forward-Deployed AI

Comfortable sitting close to users, translating feedback into product changes, and explaining technical tradeoffs clearly.

Founding AI Engineer

Has owned 0→1 builds end to end across frontend, backend, integrations, cloud deployment, and handover docs.

Hiring signal

Live deployed products over slidewareRAG, automation, and full-stack delivery in one profilePractical demos that show problem → architecture → result

Projects

Selected Work

Real, deployed builds — ordered by depth. Each has a live demo or a walkthrough so you can see how it works without running anything yourself.

01 / 03

Live demo

Voice AI Prospect Radar

A geospatial prospect-intelligence dashboard that ranks local businesses by their likely fit for Voice AI adoption, turning noisy public data into a prioritised, explainable outreach shortlist.

Problem

Teams selling Voice AI waste hours manually researching local businesses with no consistent way to judge fit, so strong leads get missed and outreach stays unfocused.

Outcome

Built an explainable scoring engine over Google Places signals with a London territory map, filterable prospect cards, and a lightweight review/ticket workflow — backed by Postgres with a reliable local-data fallback so the demo always works.

Next.jsTypeScriptPrismaSupabaseGoogle Places APITailwind

Live demo Case study GitHub

Voice AI Prospect Radar dashboard ranking London businesses by Voice AI fit

Live product — click “Live demo” to explore it yourself.

02 / 03

Live demo

RAG Chatbot System

A clean, demo-ready Retrieval-Augmented Generation chatbot with explicit ingestion, retrieval, and generation stages — it answers from your own documents, with citations and honest “I don’t know” handling.

Problem

Generic chatbots hallucinate on company-specific questions and can’t show where an answer came from, which makes them untrustworthy for real support, onboarding, and internal knowledge.

Outcome

Engineered a modular pipeline (upload → chunk → embed → retrieve → grounded answer) with source citations and explicit unknown-handling, an in-memory vector store that scales up to Pinecone, optional Azure Vision OCR for scanned PDFs, and Docker Compose for one-command demos.

TypeScriptNext.jsOpenAIPineconeAzure AI VisionDocker

Live demo GitHub

RAG Chatbot System answering a question with source citations

Video walkthrough coming soon — try the live demo in the meantime.

03 / 03

Full-stack PWA

MeetupPals

A full-stack social PWA for spontaneous local meetups (18–35) — installable and push-enabled like a native app, with profiles, map-based meetup creation, and offline-first behaviour.

Problem

Making real-life plans with people nearby is clunky across scattered chats and apps, and most web apps feel non-native on a phone and break the moment you lose signal.

Outcome

Built an installable PWA with email auth + age verification, Google Maps location selection, real-time meetup updates, push notifications, and offline-first service-worker sync — reaching a 100 Lighthouse PWA score in testing.

ReactFirebaseGoogle Maps APIStyled-ComponentsPWAService Workers

GitHub

Source on GitHub — a short demo clip will land here.

Technology

Modern Tech Stack

The tools and platforms I actually build with, day to day across AI, full-stack, and cloud. My stack leans practical over trendy — chosen so I can take a project from idea to a deployed product.

🤖

AI Coding

Cursor• AI pair programming

Windsurf• AI-first IDE

GitHub Copilot• Code completion

Claude Code• Agentic coding

v0• UI generation

🧠

AI / RAG

OpenAI API• LLMs & embeddings

Pinecone• Vector database

Vector Search• Retrieval & ranking

Azure AI Vision• OCR for scanned docs

MCP• Model Context Protocol

💻

Languages

TypeScript• Type-safe JS

JavaScript

Python• AI/ML & backend

Solidity• Ethereum smart contracts

SQL• Relational data

🛠️

Web & Backend

Next.js• Full-stack React

React

Node.js / Express• APIs & services

Tailwind CSS

Prisma• Type-safe ORM

☁️

Data & Cloud

Supabase / Postgres

Firebase• Auth, Firestore, storage

Azure• Key Vault, AI services

Docker• Containerization

Vercel• Edge deployment

Render

🚀

Delivery

Git & GitHub• Version control

CI/CD• Automated deploys

n8n / Zapier• Workflow automation

Google Maps / Places API

OAuth 2.0• Secure auth

RAG SystemsFull-StackAI Integration0-to-1 BuilderShips to Production

Expertise

Skills at a glance

A mix of tools, platforms, and collaboration muscles that teammates lean on me for—from rapid prototyping to production-ready launches.

AI & LLM Systems

RAG Systems
Vector Search & Embeddings
Prompt Engineering
Source-Cited Answers
OCR Pipelines
OpenAI API
Pinecone

Languages & Frameworks

TypeScript / JavaScript
Python
Solidity
Next.js
React
Node.js / Express

Backend, Data & Cloud

REST APIs
Firebase
Supabase / Postgres
Prisma
Azure
Docker
Vercel & CI/CD

Product & Delivery

0-to-1 Ownership
Client Discovery
Solution Design
Rapid Prototyping
Stakeholder Demos
Technical Documentation

Smart Contracts & Web3

Solidity
Ethereum Smart Contracts
Contract Testing

Currently Exploring

AI Agents & MCP
Workflow Automation (n8n / Zapier)

FAQ

Quick answers

The short version — for recruiters skimming and for AI assistants summarising my work.

Who is Suryateja Kommuri?

Suryateja Kommuri is an AI Engineer / AI Solutions Engineer based in London who builds retrieval-augmented generation (RAG) systems, workflow automation, and applied-AI products end to end — taking each from a prototype to a live, deployed URL.

What does Suryateja build?

AI products and full-stack web apps: document-grounded RAG chatbots with source citations, geospatial prospect-intelligence tools that score and prioritise leads from public data, and API/workflow automation across Next.js, TypeScript, Python, Supabase, Firebase, Azure, Docker, n8n, and Zapier. He has also written and tested Ethereum smart contracts in Solidity.

What is his strongest recent project?

Voice AI Prospect Radar — a live geospatial dashboard that ranks local businesses by their likely fit for Voice AI, with explainable scoring (Next.js, TypeScript, Prisma, Supabase, Google Places). His RAG Chatbot System is also live, with explicit ingestion → retrieval → generation stages and cited, grounded answers.

Can I try his work without installing anything?

Yes — the live demos are at voice-ai-prospect-map.vercel.app and v0-rag-chatbot-system.vercel.app. No local setup needed.

Is Suryateja available for work?

Yes. He is open to AI Engineer, AI Solutions Engineer, Forward-Deployed AI, AI Enablement, and practical full-stack AI roles in the UK or fully remote.

How can I contact him?

By email at kommurisurya@gmail.com, on LinkedIn (linkedin.com/in/suryateja-kommuri), or on GitHub (github.com/suryak02).

Let’s talk

I’m open to AI Engineer, AI Solutions Engineer, and Forward-Deployed AI roles. If you’re building with RAG, automation, or AI-assisted internal tools, tell me the workflow and I’ll show how I’d turn it into a usable shipped product.

Compose an email Say hello on LinkedIn Explore GitHub