CCTV Surveillance System
AI-powered surveillance using adversarially trained Vision Transformers for violent activity detection, custom YOLOv8 real-time detection, CLIP scene understanding, and LangChain reporting.
Muhammad Hisan Usman
AI Engineer / Research Assistant
hisan [dot] usman20 [at] gmail [dot] com
About
I am an AI Engineer leading client-facing AI engagements end to end, from discovery and solution design to delivery and executive demo, with solutions converting into six-figure contract value. Now researching agentic workflows at the University of Houston.
I am currently a Research Assistant at the University of Houston, working on agentic workflows while pursuing an MS in Engineering Data Science and AI. Previously, at Ember AI via Remotebase, I led 11 client-facing AI engagements and helped drive six-figure enterprise and healthcare contracts. At CareCloud Inc., I led a team of 10+ engineers building HIPAA-compliant voice agents and delivered an EDI claims pipeline with 95%+ accuracy.
During my BS in Artificial Intelligence, I built the CCTV Surveillance System as my final year project, combining adversarially trained Vision Transformers for violent activity detection, custom YOLOv8 object detection, CLIP scene understanding, and LangChain report generation. My current graduate work focuses on agentic workflows, alongside projects in multi-agent medical fact-checking, constitutional knowledge graphs, and travel planning with RAG.
My research interests include Agentic AI, Multi-Agent Systems, Retrieval-Augmented Generation, Natural Language Processing, Computer Vision, and Machine Learning.
Research Interests
What I work on
Technical Toolkit
What I build with
Languages & Frameworks
Python, SQL, React, PyTorch, TensorFlow, scikit-learn, LangChain, LangGraph, Transformers, FastAPI
AI & Machine Learning
Agentic AI, multi-agent systems, RAG, knowledge graphs, generative AI, LLMs, computer vision, NLP, deep learning, MLOps, data science
Cloud & Security
AWS (ECS, CloudWatch, GuardDuty, S3, RDS, Lambda), Docker, PostgreSQL, MongoDB, SOC 2, HIPAA, Vanta
Tools & Platforms
Git, GitHub Actions, Apache Airflow, DVC, MLflow, n8n, Model Context Protocol (MCP), ChromaDB, KuzuDB
Selected Projects
Systems I have built
Drug Fact-Checking Multi-Agent System
Multi-agent verification system using LLaMA 3.1, PubMedBERT, ChromaDB, and FastAPI to check drug-related queries against medical literature.
PK Constitution Knowledge Graph
Natural-language querying over Pakistan's Constitution using Qwen, Ollama, KuzuDB, and graph-based retrieval.
Multi-Agent RAG Travel Planner
Cost-optimized itinerary generation with multi-agent planning, vector embeddings, RAG, and real-time data retrieval.
Education
Academic Background
BS Artificial Intelligence
2021 – 2025
FAST-NUCES, Islamabad
Work Experience
Where I've worked
AI Engineer
Feb 2026 – Aug 2026
Ember AI via Remotebase
Junior AI Engineer
Jun 2025 – Feb 2026
CareCloud Inc.
IT Intern
Mar 2025 – May 2025
CareCloud Inc.
Machine Learning Intern
Jun 2024 – Aug 2024
Teradata
AI Intern
Jul 2023 – Sep 2023
Systems Limited
ML Intern
Jun 2023 – Jul 2023
Rapidev
Certifications
Continuous learning
Building AI Agents with RAG and LangChain
IBM · 2026
Supply Chain Planning
UC Irvine · 2025
Advanced AI: Transformers for Computer Vision
LinkedIn · 2024
Machine Learning
Stanford Online · 2022