h
hamzaziizzzz

Hamza Aziz

@hamzaziizzzz

AI Product Engineer for LLM Agents, RAG and Computer Vision

India
Inglese, Hindi
Alcune informazioni sono riportate in lingua inglese.
Chi sono
Hi, I'm Hamza - an AI Product Engineer with 3 years of production experience building LLM apps, voice agents, and computer vision systems that ship to real users. What I build: - Voice AI: Real-time DriveThru ordering agents (LangGraph, Gemini, FastAPI) - raised client success from 25% to 60%+ - LLM Apps: Full-stack RAG pipelines, voice-to-SQL converters, agentic workflows - Computer Vision: 4 facial recognition systems on DGX A100, RTX 4090, Jetson Orin Nano - 99.99% accuracy Let's build something impactful - reach out!... Continua a leggere

Competenze

h
hamzaziizzzz
Hamza Aziz
offline • 

Consulta i miei servizi

Siti Web e software IA
I will build production grade facial recognition system
A partire da75 USD / progettoMaggiori dettagli
Integrazioni IA
I will develop a real time video analytics API with tracking
A partire da100 USD / progettoMaggiori dettagli

Portfolio

Esperienza lavorativa

AI Product Engineer

VoicePlug Inc. • Full time

Aug 2025 - Present • 9 mos

Working on a production-grade DriveThru voice assistant for real-time food ordering. Contributed to NLU and conversation flow improvements within a Rasa-based voice commerce platform. Helped improve client success metrics from ~25–30% to 60%+ by addressing intent recognition, entity ambiguity, and menu understanding issues. Supported stabilization of live deployments by improving error handling, logging, and fallback behaviours. Worked closely with cross-functional teams (product, support, QA) to iterate on conversational quality based on real user interactions.

Assistant AI Engineer

Global Infoventures Pvt. Lyd. • Full time

Jul 2023 - Aug 2025 • 2 yrs 1 mo

Led development of 4 commercial-grade Facial Recognition Attendance Systems on DGX A100, RTX 4090, Jetson Orin Nano, achieving >99.99% accuracy Developed G6 Voice Assistant, converting natural language voice queries into precise SQL queries via Rasa, Gemini 2.0, and Milvus Vector DB; handled ambiguous queries with advanced intent disambiguation, achieving near-100% intent classification accuracy Mitigated domain shift in facial recognition using a CI/CD-style pipeline combining mobile and CCTV datasets — improved cross-domain recognition by 14% Automated nightly face vector updates from 8ft-high CCTV, boosting real-world accuracy from 85% to 99% Integrated PostgreSQL for recognition logs and built an API sync module with legacy SIM ERP Portal, enabling real-time attendance updates and SMS alerts for over 2,000+ students, enhancing parent engagement and system transparency