Hi, I'm
Senior Full-Stack
AI Engineer
HamzaQureshi
I architect production AI agents & LLM systems that actually ship.
Download CVWho I am
Anyone can demo AI. Few can keep it running in production.
I'm Hamza Qureshi, a Senior Full-Stack AI Engineer with 5 years shipping production agentic systems. As team lead at Evren AI, I architect LangGraph multi-agent platforms for US clients, from a live workplace-mediation product to a HIPAA-aware financial assistant. I turn LLMs, RAG, and real-time voice into systems people actually rely on.
Capabilities
What I build.
From AI systems to the infrastructure they run on, end to end.
AI Agents & Multi-Agent Systems
Stateful, orchestrated LangGraph multi-agent systems that reason, route, and act, with crisis-detection and anti-hallucination safety layers.
RAG & LLM Engineering
Retrieval-augmented generation with pgvector semantic search, prompt engineering, and model selection across GPT-4o, Claude, and OpenRouter.
Voice & Real-Time AI
Sub-2-second voice pipelines that stream STT → LLM → TTS with interruption support, over WebSockets and SSE.
Full-Stack & Cloud
End-to-end delivery of typed FastAPI backends and Next.js frontends, deployed on AWS with Docker and CI/CD.
Computer Vision & ML
Real-time facial recognition and monitoring with MediaPipe quality gating and dlib embeddings, at 90%+ live accuracy.
Experience
The journey so far.
Evren AI
2024, PresentLead AI Engineer
- Technical team lead architecting LangGraph agentic systems for US clients, including Common Ground, a live AI workplace-mediation pilot, and Jarvis, a HIPAA-aware multi-agent financial assistant integrating 5 healthcare platforms.
- Architected Converscraft's real-time voice pipeline (Deepgram → GPT-4o → ElevenLabs) and broke the latency ceiling to sub-2-second end-to-end responses.
- Lead and mentor a cross-functional team of 6 engineers, owning architecture, code reviews, sprint planning, and US client communication.
DevHexa
2024, PresentSenior AI Engineer
- Independently architect production AI systems for European and US clients: stateful multi-agent LangChain pipelines, RAG with pgvector semantic search, and GPT-4o FastAPI backends.
- Ship full-stack platforms end-to-end on AWS, owning every layer from schema to production: system design, API architecture, Next.js frontends, Docker, and CI/CD.
- Benchmark and select LLMs across OpenAI, Anthropic, and OpenRouter, owning prompt strategy and the cost-latency tradeoffs that shape the deployment stack.
Vareya
2023, 2024Full-Stack AI Engineer
- Delivered GPT-powered SaaS apps and production REST APIs for international clients, progressing from frontend to owning complete AI product development within 6 months.
- Built full-stack apps on AWS with FastAPI, PostgreSQL, and Docker, integrating early LangChain workflows and OpenAI APIs as agentic tooling matured.
- Automated complex business workflows with N8N and Make.com and maintained CI/CD via GitHub Actions across concurrent engagements.
Freelance
2021, 2022Full-Stack Developer
- Built and delivered web applications for small-business clients end-to-end: Next.js/React frontends, Python/Node.js backends, database design, and cloud deployment.
- Developed Python automation tools and data-processing workflows that laid the foundation for the production AI engineering work that followed.
Selected Work
Real systems, real users.
A closer look at the products and AI systems I've shipped.
AI Mediation Platform for Employee Relations
An AI platform for workplace case resolution across conflict, raise requests, and performance reviews. I led the LangGraph agent architecture, memory system, routing logic, and question-generation flow that power structured, multi-turn conversations with reliable handoff between case types. It is live in pilot with real organizational users.
Real-Time AI Language Learning Platform
A real-time AI language-learning platform for German, Spanish, French, and English. I led the voice AI architecture across speech-to-text, LLM, text-to-speech, WebSocket infrastructure, and interruption handling. It hits sub-2-second end-to-end latency for a natural conversational experience, and powers quizzes, streaks, and module progression in production.
AI-Powered Financial Platform for Healthcare Providers
A HIPAA-compliant AI financial platform for healthcare providers, built on a LangGraph multi-agent architecture. I designed four specialized ReAct agents (Supervisor, Financial, Clinical, Marketing), each with dedicated toolsets and integrations like Sage, Bill.com, HubSpot, and SharePoint. Built on FastAPI, PostgreSQL, and AWS with RPA and ETL pipelines, it delivers an AI voice and chat assistant, cash-flow analytics, P&L reporting, and intelligent document search, with full RBAC, audit logging, and encrypted data handling.
FacialTrack: Facial Recognition Attendance System
A real-time AI attendance system using facial recognition for automated classroom monitoring. I engineered a Python backend with OpenCV and FaceNet for live face detection from webcam feeds, a PostgreSQL schema for session logs and entry/exit tracking, and RESTful APIs for a Flutter Android app with role-based access for multi-teacher, multi-subject management.
AI-Powered Sales Talent Matching Platform
A full-stack AI hiring platform for sales recruitment that replaces traditional résumés with verified performance data. I engineered an AI sales-fit score using ML models and OpenAI across 11 metrics, built FastAPI scoring microservices on AWS, and integrated Apollo.io enrichment with a Next.js recruiter dashboard. The platform served 12,000+ users, supported 450+ placements, and cut hiring cycles by 50%.
AI Business Incubator Platform
An AI-powered business-validation platform that takes users from raw ideas to structured market research in one workflow. I owned the full product solo, from the frontend and backend to the database and LangChain orchestration across Perplexity and OpenAI. It handles persona creation, pain-point discovery from real web and Reddit signals, idea generation, and long-form research reports, cutting early-stage validation from weeks to minutes. Deployed on AWS.
AI Personal Operations Agent
A multi-agent AI assistant that automates Gmail, Google Calendar, and GitHub. I owned the full LangGraph architecture across planning, execution, semantic memory, voice input, and LangSmith tracing. It demonstrates real agentic workflows, persistent context, and production-style integrations across everyday productivity tools.
AI Platform for Healthcare Intelligence & Workflow Management
An AI platform for healthcare intelligence and workflow management that helps teams handle complex information, structured processes, and operational workflows. I owned the product end to end as the solo engineer, from application architecture and backend systems to data flow and the AI integration layer. It is designed for secure healthcare operations and intelligent data handling that feels practical and product-ready, not experimental.
Get in touch
Let's build something exceptional.
Open to senior AI engineering roles, consulting, and select client builds. Have an agentic system to ship or a team to lead? Drop a message, or email me directly.