Services Work Process Contact + Start a Project
Hassan
X:0 Y:0
I ship
AI that
works.
AI Engineer  ·  Freelance  ·  Open to Projects
Hackathon Winner
8+
Deployed Projects
<48h
Time to First Demo
Scroll
Available Now
AI systems, automation pipelines, and full-stack products — built and deployed, not just designed.
AI Automation
On-chain Attestation
RAG Pipelines
Multi-Agent Systems
Satellite ML
Full-Stack Products
LLMs in Production
Web3 Infrastructure
AI Automation
On-chain Attestation
RAG Pipelines
Multi-Agent Systems
Satellite ML
Full-Stack Products
LLMs in Production
Web3 Infrastructure
What I Build

AI products that ship.

01 — Automation

AI Pipelines & Agents

Multi-agent workflows, RAG systems, and LLM pipelines that run in production — not in notebooks. From data ingestion to delivery, fully deployed.

02 — Full-Stack

Products & Platforms

End-to-end web applications with AI at the core. Next.js frontends, FastAPI backends, deployed and monitored from day one. No handoff zip files.

03 — Web3 + AI

On-chain Intelligence

Smart contracts with AI integration — attestation systems, verifiable computation, and agent memory with blockchain-level guarantees built in.

Hackathon Winner
8+
Shipped Systems
<48h
Time to First Demo
6/6
Production Validated
selected work
IEEE Codebhoomi — Top 10 National
fumii — AI Companion Desktop App
Cross-platform Electron app with tiered RAG memory, multi-provider LLM abstraction, and a 9-state sentiment-driven animation system.
ElectronTypeScriptSQLiteMistralOpenAIClaude
↗ GitHub
IIT Bombay Eureka — Semi-Finalist
DeTalks — Multi-Agent Mental Wellness
Dual-agent platform with hard safety architecture. "Disha" triages. "Kavach" silently monitors and generates warm-handoff summaries for therapists.
ReactTypeScriptViteshadcn/uiTailwind
↗ GitHub
River Watch — Satellite Anomaly Detection
Dual-signal satellite fusion for detecting illegal sand mining anomalies across Indian river systems. Sentinel-1 SAR + Sentinel-2 NDWI differencing.
PythonEarth EngineSentinel-1 SARStreamlit
↗ GitHub
Mike — AI Voice Dictation
Under 500ms transcription latency via Groq Whisper, three LLM polishing modes, custom RMS voice-activity detection. Single-executable Windows build.
PythonPyQt6Groq WhisperPyInstaller
↗ GitHub
EvrythingAI — Newsletter Pipeline
Zero-infrastructure production LLM pipeline: RSS ingestion → multi-document summarization → synthesis → timezone-aware delivery at subscribers' local 6am. Cost: $0/month.
Node.jsMistral AIGitHub ActionsResend
↗ GitHub
1ph — One Place for Hackathons
Hackathon aggregator with AI enrichment pipeline (Mistral AI) that extracts real deadlines, prize pools, and descriptions from raw unstructured listings.
Next.js 14PrismaMistral AIGitHub Actions
↗ GitHub
Process

From idea to deployed.

01
Scope

You bring the problem. I define what's technically possible, what's worth building, and what the MVP looks like. One call, no fluff.

02
Build

I ship in tight cycles. Working demos within days, not weeks. You see progress constantly — no black boxes, no surprise pivots.

03
Deploy

Production-ready from day one. CI/CD, monitoring, error tracking. I hand off a live system with docs — not a zip file.

04
Iterate

Post-launch support included. If something breaks or needs improving after go-live, I'm already there — no extra retainer needed.

Tech I use
PythonTypeScriptJavaScriptSQLSolidityNext.js 14ReactFastAPIFastifyNode.jsElectronStreamlitOpenAIClaude APIMistralGroq / WhisperRAG / pgvectorMulti-agent LLMsLLM Orchestrationscikit-learnLightGBMPandasNumPyxarrayGoogle Earth EngineSentinel-2NASA GPMSolidityHardhatOpenZeppelinviemMCP ServerDockerRedisPostgreSQLNeo4jCeleryPrometheusNginxGitHub ActionsVercelPytestTurborepo

Let's build
something
real.

Have a problem that needs AI, automation, or full-stack engineering? I take on projects where the technical depth matters. Tell me what you're building.