Open to Remote AI / ML Opportunities

Muhammad
Haseeb
Ashraf

AI/ML Engineer focused on building production-ready AI systems across Computer Vision, Generative AI, Vision-Language Models, Edge AI, and scalable AI infrastructure.

Computer Vision LLMs VLMs RAG Edge AI Qualcomm AI Satellite AI Backend Systems
Get In Touch LinkedIn GitHub
Muhammad Haseeb Ashraf

AI/ML Engineer

Building reliable AI systems that solve real-world problems.

3.73

CGPA — FAST-NUCES

8×

Dean's List

5+

International Clients

2×

ICPC Qualifier

About

Building AI That Actually Ships

I enjoy working on difficult engineering problems where AI needs to move beyond experimentation and operate reliably in production environments. My work spans computer vision, edge AI, generative AI systems, satellite imagery, and multimodal intelligence.

I specialize in designing production-ready AI systems with a strong focus on performance, scalability, and real-world deployment.

Over the last few years, I have worked across healthcare AI, satellite imagery, conversational AI, OCR systems, document intelligence, and on-device AI inference.

My engineering style focuses on ownership, fast iteration, deployment-oriented thinking, and building systems that create measurable impact instead of staying limited to research notebooks.

170K+

Sq km imagery processed in geospatial AI pipelines

95%

Reduction in manual extraction via OCR automation systems

100+

Conversation-turn evaluation pipeline for AI voice agents

Experience

Professional Experience

AI / ML Engineer

April 2025 — Present
Edge C Pvt. Ltd. · Lahore, Pakistan
  • Built large-scale SAR-based geospatial intelligence pipelines using multi-polarization satellite data for industrial asset detection in remote environments, enabling automated monitoring where traditional field surveys were slow, costly, and weather-dependent.
  • Developed and optimized production-grade YOLO11-based detection systems for aerial and satellite imagery analysis of large-scale infrastructure, supporting high-throughput automated inspection workflows with consistent real-world accuracy.
  • Engineered a geospatial intelligence pipeline combining multi-source satellite imagery, aerial datasets, object detection models, and spatial data enrichment techniques to assess large-area infrastructure characteristics, reducing manual analysis cycles from weeks to automated compute-driven workflows.
  • Optimized deep learning inference pipelines (LightGlue and vision models) for Qualcomm Snapdragon hardware, enabling hybrid CPU/GPU/NPU execution for offline and bandwidth-constrained edge environments.
  • Developed OCR and document intelligence systems using PaddleOCR and DeepSeek OCR for structured extraction from complex documents, reducing manual processing workload by 95%+.
  • Built conversational AI and voice-driven automation systems with evaluation pipelines to improve response quality, reliability, and scalability in production environments.

AI Research Assistant

Feb 2024 — Sep 2025
FAST-NUCES · Lahore, Pakistan
  • Conducted interpretability analysis of transformer-based LLM architectures using Explainable AI techniques to study model reasoning behavior and performance tradeoffs.
  • Designed a secure access control framework for distributed IoT environments using blockchain-backed policy enforcement and game-theoretic decision models.
  • Developed NLP pipelines for toxic content detection in large-scale online communities using transformer-based classification and topic modeling techniques.
Skills

Technical Stack

AI / ML

PyTorch TensorFlow Scikit-learn Transformers Fine-tuning LoRA / PEFT

LLMs & GenAI

RAG Graph RAG LangChain LangGraph LlamaIndex OpenAI API

Computer Vision

YOLO v8-v11 OpenCV OCR Satellite AI SAR Imagery BLIP-2

Edge AI

Qualcomm AI Toolkit TensorRT ONNX Model Quantization GPU/NPU Optimization

Backend & Cloud

Python FastAPI Docker Kubernetes AWS SageMaker Azure AI

Infrastructure

CI/CD FAISS ChromaDB Pinecone LangFuse
Projects

Selected Projects

US Energy

Energy Asset Detection System

Production SAR imagery classifier for oil rig and frac site detection enabling automated surveillance across remote energy territories.

YOLO SAR CVAT Python
Geospatial AI

Geospatial Infrastructure Intelligence Pipeline

Enterprise AI pipeline processing MAXAR and VEXCEL imagery for automated road infrastructure assessment at scale.

YOLO QGIS OSM MAXAR
Fintech

Enterprise Document Intelligence Platform

Graph RAG retrieval system across 1M+ financial documents with relationship-aware semantic search.

Neo4j FastAPI Graph RAG Python
Healthcare AI

Medical Consultation Intelligence Platform

AI healthcare platform transcribing doctor-patient conversations and generating clinical recommendations.

Next.js LlamaIndex MongoDB OpenAI
Conversational AI

Multi-Domain Conversational AI System

100+ module conversational AI platform for media research and production workflows.

Django LangChain LLMs
Edge AI

On-Device YOLO Android Application

Fully on-device real-time object detection app running YOLO26n natively on mid-range Android hardware.

Android YOLO26n Qualcomm AI
Education

Academic Background

BSc Computer Science

FAST-National University of Computer and Emerging Sciences

CGPA: 3.73 / 4.0 · QS Rank #651–700


Achievements:

  • Cum Laude Honors
  • Dean’s List (8 Consecutive Terms)
  • Bronze Medal ×2
  • ICPC Qualifier 2021 & 2025
  • Leading Edge Skills American Scholarship

Let's Build Something Meaningful

Open to remote AI engineering opportunities, research collaborations, and ambitious projects involving Computer Vision, Generative AI, Edge AI, and scalable intelligent systems.

im-haseeb@outlook.com LinkedIn GitHub