Open to opportunities

Maowiz Saleem

@maowizsaleem

Machine Learning Engineer specializing in computer vision and generative AI—shipping production-ready end-to-end ML systems.

Pakistan

Message

What I'm looking for

I’m looking for a role where I can build and deploy production ML systems—especially computer vision and RAG—owning end-to-end delivery (APIs, containers, and serving) and collaborating to turn models into measurable business impact.

I’m a Machine Learning Engineer with 2+ years of hands-on experience designing, training, and deploying end-to-end ML and computer vision systems for clients across UAE, USA, and Pakistan.

I’ve shipped real-world solutions across the full ML lifecycle—data collection, feature engineering, model training and evaluation, REST API development, and containerised deployment with Docker—so models reliably reach production.

Highlights include a real-time PPE compliance detector (∼91% mAP) that reduced manual inspection by ∼60%, a voice AI calling platform that cut call-handling time by ∼3 minutes per call, and a demand-forecasting platform processing 3M+ transactions with ∼18% RMSE improvement. I also build RAG-powered knowledge bots and privacy-first assistants optimized for low-resource, CPU-only inference using llama.cpp.

Experience

Work history, roles, and key accomplishments

Current

AI/ML Consultant

Current

Independent AI/ML Consultant

May 2024 - Present (2 years 1 month)

Engineered a real-time PPE compliance vision system (YOLO/OpenCV) achieving ~91% mAP and reducing on-site manual inspection overhead by ~60%. Built an automated voice calling platform and an offline RAG knowledge pipeline using embeddings and local LLM inference, delivering ~87% answer relevance on test queries.

YOLOv8 OpenCV fastAPI Retrieval Augmented Generation (RAG)Embeddings Faiss Docker Prompt Engineering

Machine Learning Research Assistant

Comsats University Islamabad

Jan 2024 - Jan 2026 (2 years)

Curated and annotated a 12,000+ image dataset for supervised learning research and improved baseline image classification accuracy by ~8% through data augmentation and hyperparameter tuning. Documented dataset preparation and experimental workflows to support research reproducibility.

Hyperparameter Tuning

Education

Degrees, certifications, and relevant coursework

COMSATS University Islamabad

Bachelor of Science in Software Engineering, Software Engineering

2022 - 2026

Grade: CGPA: 3.95/4.0 (Gold Medalist)

Activities and societies: Relevant coursework: Machine Learning, Artificial Intelligence, Deep Learning, Data Structures & Algorithms, Database Systems, Computer Vision, Linear Algebra, Operating Systems.

Completed a B.Sc. in Software Engineering (Jan 2022–Jan 2026) with CGPA 3.95/4.0 and Gold Medalist recognition. Coursework covered machine learning, artificial intelligence, deep learning, computer vision, and core CS foundations.