Lahore, Pakistan · Open to AI engineering roles
Abdul Rehman
AI Engineer & Published Researcher
AI Engineer building agentic AI, RAG systems, and LLM-powered backends, with a research background in deep learning and computer vision. I focus on the hard part of AI engineering: making models dependable on messy, real-world data.

Get to know me
About Me
I'm an AI Engineer working on agentic AI, retrieval-augmented generation (RAG), and LLM-powered backends, with a research background in deep learning and computer vision. I enjoy the hard part of AI engineering: making models dependable on messy, real-world data.
At Nextbridge, I build agentic AI and RAG systems for real business use cases. At Digitalytics, I design AI automation pipelines and voice agents that streamline operational workflows.
Before industry, I spent nearly three years as a research member at VLCMatrix Lab, working on deep learning for medical imaging, deepfake detection, and agriculture. That work led to peer-reviewed publications in IEEE, Springer, and Wiley, and gave me a habit I still rely on: test carefully, measure honestly, and understand why a model works before shipping it.
0
Published Papers
0
Class Rank
0
CGPA
0
Certifications
Published in
Where I've worked
Work Experience
AI Engineer
Nextbridge Ltd.
- ▹Engineering LLM systems with LangGraph and FastAPI on self-hosted Ollama models, including a text-to-SQL database agent, a Graph RAG pipeline on Neo4j, a hybrid-search RAG chatbot on Qdrant, and a multi-agent coding assistant with a Docker sandbox.
- ▹Trained computer vision models for PPE compliance detection on construction sites.
- ▹Built a multi-person pose estimation pipeline that tracks consistent person IDs across video frames.
AI Automation Engineer (Part-time)
Digitalytics
- ▹Building AI-powered EHR automation pipelines that extract, validate, and enter structured data from documents into web-based systems.
- ▹Deploying AI voice agents on Retell AI with Twilio telephony.
- ▹Building n8n workflows that connect LLMs with third-party APIs.
Research Lab Member
VLCMatrix Lab, University of Lahore
- ▹Spent nearly three years on deep learning for medical imaging, deepfake detection, and agriculture.
- ▹Managed end-to-end machine learning workflows: data preprocessing, model training, hyperparameter tuning, and evaluation.
- ▹Research contributed to 8 peer-reviewed papers published across IEEE, Springer, and Wiley.
What I work with
Skills & Technologies
AI & Agentic AI
ML & Deep Learning
Programming Languages
Web & Backend
Data & APIs
DevOps & Tools
What I've built
Projects
> ask "top 10 customers by revenue"
[LLM] Parsing question...
[SCHEMA] 4 tables retrieved (Qdrant)
[SQL] Compiled via SQLGlot
[DONE] 12 rows, chart rendered
█Featured Project
Database Agent
A LangGraph database agent that turns natural-language questions into governed, safe SQL over Postgres, MySQL, CSV, and Excel. It builds an LLM-generated YAML semantic layer, compiles queries with SQLGlot, retrieves schema context from Qdrant, and auto-generates charts from the results.
> query: "What is our leave policy?"
[RETRIEVE] Hybrid dense + BM25...
[RERANK] FlashRank top-5 selected
[ANSWER] Grounded response ready
[VOICE] Piper TTS streaming...
█Featured Project
RAG Assistant
An employee RAG assistant with hybrid dense and BM25 retrieval in Qdrant, FlashRank reranking, and a web-search fallback. It automates HR requests through the Gmail API, keeps Postgres-backed multi-turn memory, and adds voice interaction with Whisper and Piper.
> python ingest.py report.pdf
[EXTRACT] Entities + relations (LLM)
[GRAPH] Writing nodes to Neo4j
[DEDUP] Embedding-based merge done
[QUERY] Cypher retrieval ready
█Featured Project
Graph RAG
A Graph RAG pipeline that converts PDF documents into a Neo4j knowledge graph using LLM entity extraction and embedding-based deduplication. It runs multi-strategy Cypher retrieval with query-intent filtering and result ranking to answer questions, served locally on Ollama.
> python surveillance.py
[INFO] YOLO11 model loaded
[ALERT] Knife detected (0.94)
[TTS] Speaking alert...
[INFO] Dashboard SSE active
█Featured Project
YOLO11 Intelligent Surveillance System
Real-time AI surveillance deployed on Raspberry Pi that detects fire, knives, guns, cigarettes, and vapes. It uses temporal verification to prevent false-alarm spam, GPT-4o Vision as a second-layer false-positive reducer, and Piper TTS for voice alerts. Ships a live web dashboard with real-time charts, event history, and start/stop controls via SSE.
Other Noteworthy Projects
Real Estate Backend
A FastAPI platform with JWT role-based access, lead management, WebSocket updates, signed webhooks, and a tool-calling AI agent, load-tested with pytest, Locust, and Newman.
LangGraph AppBuilder
A multi-agent LangGraph system where Planner, Architect, and Coder agents collaborate to generate full working applications from a single prompt.
Coding Agent
A modular coding agent built with LangChain and Ollama that writes, edits, and executes code inside a Docker sandbox.
Research output
Publications
8 peer-reviewed papers as an undergraduate across IEEE, Wiley, and Springer, covering agentic AI evaluation, AI surveillance, deepfake detection, medical imaging, satellite imagery analysis, autonomous vehicle platooning, and agricultural AI.
★ Journal Publication
From Benchmarks to Deployment: A Comprehensive Review of Agentic AI Evaluation
Artificial Intelligence Review · 2026
AI-Powered Surveillance Systems: A Comprehensive Review of Technologies, Applications, Challenges, and Deployment Considerations
WIREs Data Mining and Knowledge Discovery · 2026
Enhancing Deepfake Video Detection: An Xception-CBAM-BiLSTM Framework with Attention Mechanisms
ICETCS 2025 · 2026
CADSNet: Cross-Aligned Dual-Stream Network for Building Damage Assessment Using Disaster Satellite Imagery
ICCA 2025 · 2025
A Multi-Agent Reinforcement Learning Approach for Seamless and Safe Autonomous Vehicle Platooning
Intl. Conf. on Intelligent Technology, Systems & Service · 2025
Data-Efficient Wheat Disease Detection using Shifted Window Transformer: Enhancing Accuracy, Sustainability, and Global Food Security
IEEE Transactions on Consumer Electronics · 2025
Enhanced Model for Sign Language Detection with Deep CNN: Achieving Perfect Accuracy in Hand Gesture Recognition
ICIC 2024 · 2024
Optimizing Diabetic Foot Ulcer Detection: Leveraging SSD and YOLO Architectures
HONET 2024 · 2024
Academic background
Education
Bachelor of Science in Computer Science (BSCS)
University of Lahore, Sargodha Campus
Achieved 2nd-in-class ranking while simultaneously authoring 8 peer-reviewed publications across IEEE, Springer, and Wiley as an undergraduate.
▹ Certifications
Generative AI Applications Specialist
Fundamentals of AI Agents Using RAG and LangChain
Project: Generative AI Applications with RAG and LangChain
Generative AI: Prompt Engineering Basics
Level 2 Fundamentals of Agentic AI Professional Certification
Agent Factory Fundamentals: Building Digital Full-Time Equivalents (FTEs)
Fundamentals of Agentic AI Level 1
Fundamentals of Modern AI Python Level 1
Prompt and Context Engineering: Effective AI Communication Level 1
What's Next?
Let's Build Something
I'm open to new roles, freelance projects, and research collaborations. If you have something in mind, or just want to say hi, my inbox is always open.