Altaf Ur Rahman

About

Senior AI engineer with 8+ years of experience building, deploying, and leading AI solutions across computer vision, classical machine learning, and agentic AI systems. Experienced delivering production systems on GCP and AWS for real-time users across the MENA region.

Experience

• Senior AI Expert | Consultant
  Tech Horizons Ventures · Riyadh
  Sep 2025 – Apr 2026

  Worked as AI Consultant to build and deploy producttion grade Agentic AI products for Ghorfatech, Rahatech, Seyaha, Decotis etc. The products are live and serving users in different industries. The products were mapped from product requirements, developing, integrating and deploying in GCP instances to serve them live. Technologies ranged from Python, LLMs, Rest API's, RAG, Agentic Workflows, GCP deployements etc.

• AI/ML Expert (AI Team Lead)
  Green AI · Islamabad, Pakistan
  Sep 2024 – Nov 2025

  My role here was a lead resouce person as AI Expert to incorporate, monitor and develop the Artificial Intelligence models for precision agriculture in this government project including Ministry of Defence, Punjab Information Technology Board and National Aerospace(PITB), Science and Technoloy Park(NASTP). Green AI's Agriverse maps Pakistani indiginous different crops with technologies such Artificial Intelligence, Cloud with data layers from Satellite, drone, Mobile images, IOTs etc. All this Machine Learning work was designed into 24+ models whose ultimate goal was to monitor and grow yield of Pakistani crops benefiting different users such as government, industries, farmers, institutions etc. Tech stack was diverse since we have developed alot of models including classical ML such as using Regresssion, XGBoost, NNs, etc with computer vision models such as YOLO. We have regorously done A/B testing with different algorithms in order to maximize performance measures with PITB and in house team. Few of the models were [Satellite data: Sentinal] Crop Type Detection, Crop Stage detection, Crop Growth trend Analysis, Crop health Analysis, Soil Segmentation, Water Content Indication, Crop Yield detection, Crop Budgeting and Revenue Optimizations etc. Models were integrated into Agriverse and deployed on own cloud. Here are the Responsibilities and Outcomes produced under my guidance: - Led and Monitored the different AI models for the modules including Vegetation Change Monitoring and Prediction{VCMAP}, Crop disease and Pest Detection (CD&PD) and Soild Quality Analysis (SQA) with PITB and in-house Team. - Developed Evals and Checklists to evaluate models developed by PITB, integration and deployement compliance and value deliverables as per Software Quality Specifications (SRS) - Delegating updates, Syncs, Meeting Requirements between PITB and PMU Green AI - Improved the Crop detection accuracy from 15% to 89% with satellite data for Pakistani Cro

• Machine Learning Engineer
  AUG AI
  Nov 2022 – Sep 2024

  Developed and deployed multiple AI models including legalassistant.au and US Banks fraudDetector using LLMs and GNNs on AWS. Hit 100% recall and 95% precision on a heavily imbalanced dataset. [Team of 5]

• Lecturer
  Air University
  Jul 2018 – Sep 2022

  Built, supervised and taught AI and software applications in engineering systems including machine learning, computer vision, Robotics and actuating systems and its integration Engineering.

• Research Assistant
  NUST
  Sep 2017 – Jun 2018

  Researched computer vision and NLP for Urdu text recognition including image processing and using VGG-16 to extract text regions.

Education

• M.Sc
  School of Mechanical and Manufacturing Engineering · National Unviersity of Sciences and Technology, Islamabad
  Robotics and Intelligent Machines Engineering · Sep 2015 – Aug 2017
• BS Mechatronics Engineering
  University of Engineering and Technology (UET), Peshawar
  Mechatronics Engineering · Sep 2011 – Aug 2015

Skills

• C
• Google Cloud Platforms (GCP)
• Python
• scikit-learn
• PyTorch
• LLM evaluation
• LangChain
• Yolo
• Natural Language Processing
• Observability
• RESTful APIs
• OCR
• RAG
• AIOps
• Vertex AI
• Document Intelligence
• Context Engineering
• Agentic Engineering

Projects

• Detection and tracking of objects in Two Phase flow Air-Water Channel
  Image Processing, Kalman Filter

  Tracked bubbles and their velocities in nonlinear channels for an oil and gas application, improving results by up to 25%.

• Fraud Detection in Loan Applications
  LLM, GNN, AWS, Python, XGBoost, Experiment Tracking, CI/CD, GitHub Actions

  Fraud detection for US bank loan applications using LLMs and graph neural networks, deployed on AWS.

• Hospital Conversational Agent
  Graph RAG, Document Ingestion, Groq, LangChain, FAISS, SentenceTransformers, Neo4j, Python

  Conversational AI agent that retrieves healthcare information using knowledge graphs and NLP to query a Neo4j database.

• Ghorfatech Assistant ↗
  LangGraph, Prompt Engineering, Context Engineering, Redis, AI Guardrails, GCP, HITL

  Agentic AI assistant for Saudi Arabia’s vacation rental marketplace with intent-aware RAG, role-based tool invocation, and human escalation handling.

• Rahatech Agentic System ↗
  Agent Harness, Context Engineering, Cache management, AI Security, MCP, GCP, HITL

  Multi-agent mental health and counseling platform with agentic tool-calling, agent router, low-latency search, and response-improvement loop engineering.

Courses & certifications

• LangChain for LLM Application Development
• Generative AI with Large Language Models
• Deep Learning Specialization · Coursera / Andrew Ng
• Machine Learning Specialization · Coursera / Andrew Ng
• Agentic AI and AI Agents

📚 Publications

• Detection and tracking of bubbles in two-phase air water flow for non-convergent sinusoidal channel · 2018

🎯 Hobbies & interests

• Marathon
• Running
• Hiking