@mfaizanq
Building AI systems from LLMs and RAG pipelines to real-world applications.
// ABOUT
Production-focused ML systems that connect research to real business outcomes.
1-- Oracle Agile PLM · prompt-to-SQL context2SELECT av.item_number, av.description3FROM agile_item av4JOIN agile_rev ar ON av.id = ar.item5WHERE av.subclass = :cls6 AND ar.rev_number = (7 SELECT MAX(rev_number) FROM agile_rev WHERE item = av.id8 );
I design and build production-grade AI systems, combining LLMs, retrieval pipelines, and scalable backend architectures to solve real-world problems.
My experience spans generative AI, computer vision, and full-stack development, allowing me to build complete, end-to-end intelligent systems.
From RAG pipelines to YOLO-based detection systems and diffusion model fine-tuning, I focus on building solutions that are not just functional but reliable and scalable.
900+
LLM Fine-Tuning Q&A Pairs
3+
Years Building AI Systems
5+
Production AI Projects
10+
AI Tools & Frameworks Used
// PROJECTS
Flagship work first — then a broader set of shipped systems and experiments.
Featured
Larger-impact work — production deploys, proprietary models, and products live in the wild.
Fine-tuned GPT OSS-20B using LoRA on Oracle Agile schemas for schema understanding and prompt-to-SQL generation using 900+ curated Q&A pairs.
AI-driven system for extracting, validating, and analyzing audit data from structured and unstructured documents.
More work
Shorter summaries — same rigor, scoped for a quick scan.
Computer vision system that highlights differences between CAD revisions to speed up design validation.
Built a local RAG system using ChromaDB, LlamaIndex, and Mistral for secure legal document search.
Built enterprise AI agents with LLM integration, custom workflows, and secure connectivity to business data.
Developed dashboards and reports using complex SQL with optimized layouts for reliable stakeholder insights.
PyTorch-based multi-output model for 5D CFD simulations using custom loss functions and tuning for accuracy.
Real-time player and ball tracking system using YOLOv11, OpenCV, and a custom-trained dataset.
Fine-tuned Stable Diffusion v1.5 for generating consistent, brand-specific product images.
// SKILLS
The tools and technologies I use to ship end-to-end AI and data products.
// EXPERIENCE
From applied computer vision and RAG to enterprise Oracle Cloud AI.
AI and Data Engineer (ASE)
AI and Data Engineer
Project Intern
// EDUCATION
Foundations in computer science from NUST and secondary qualifications.
Islamabad, Pakistan
Bachelor's in Computer Science
CGPA 3.23 · 2021 — 2025
Lahore, Pakistan
A-levels (1A* 3A)
2019 — 2021
Lahore, Pakistan
O-levels (6A* 4A)
2017 — 2019
// CONTACT
Have a project in mind? I'm always open to discussing new opportunities.