Hi, I'm Zaraar Malik
Machine Learning Engineer
I create end-to-end machine learning pipelines that power smarter products and better decisions.
About Me
Get to know me better
I'm a Machine Learning Engineer passionate about AI, deep learning, and building production-ready intelligent systems. I enjoy transforming ideas into real-world AI solutions.
I specialize in scalable ML models, end-to-end pipelines, and GenAI applications. My work spans LLMs, RAG systems, computer vision, and small language models, always exploring new AI frontiers.
Outside of coding, I research emerging study about AI trends, study about new technologies.
Work Experience
My professional journey
Machine Learning Engineer
Apexbeat - Manchester - United Kingdom
September 2025 - Present
- ▸Built a production-grade multi-modal medical chatbot.
- ▸Engineered automated pipelines for document ingestion, formatting, cleaning and DB insertion.
- ▸Deployed MMM-RAG chatbot on AWS handling for 2,000+ active users.
- ▸Integrated MongoDB for chat histories and personalized user interaction data.
AI Intern
Systems Limited - Islamabad - Pakistan
June 2025 - August 2025
- ▸Fine-tuned transformer LLMs for internal use cases to improve latency and relevance.
- ▸Built Azure pipelines and containerized AI services using Docker.
- ▸Ran feasibility studies on GenAI projects, assessing trade-offs and performance.
- ▸Optimized model serving for better cost, speed, and reliability.
AI Research Assistant
FSM - Islamabad - Pakistan
September 2024 - March 2025
- ▸Enhanced financial ML datasets in collaboration with Data Engineering.
- ▸Ran churn, credit risk, and investment likelihood experiments.
- ▸Built a RAG-based Investment Assistant with region-specific insights.
- ▸Developed reproducible pipelines for fast model testing and iteration.
AI Intern
AIO (Silicon Valley) - Islamabad - Pakistan
June 2024 - August 2024
- ▸Worked with Data Engineering team to refine and improve data for model training
- ▸Developed Custom Evaluation Metrics for the model
- ▸Optimized model performance by 20% through hyperparameter tuning and quantization
Skills & Technologies
Technologies I work with
Programming & Frontend
AI & Deep Learning
Databases & Pipelines
Tools & Deployment
Featured Projects
Some of my recent work
Multi-Agent Chatbot for Medical Reimbursement
AI assistant automating medical receipt reimbursements, enabling agent collaboration, validation checks, and error handling to reduce manual effort.
TinyLLM RAG Chatbot for Task-Specific Queries
Experimented with lightweight LLMs for RAG, optimizing chunking strategies and evaluating trade-offs between efficiency, accuracy, and deployment cost.
Multi-Document RAG System
Developed a system to retrieve and synthesize insights across large document collections, designing chunking strategies, indexing pipelines, and scalable retrieval methods.
Snap Shop – GenAI Fashion Synthesizer
Web-based try-on platform using LoRA-tuned diffusion models for garment synthesis and fashion visualization.
Procedural Game Level Generation
Generated playable Mario levels using DCGAN and WGAN architectures to automate game design workflows.
Hidden Object Detection
Fine-tuned YOLOv8 and Faster R-CNN on iMaterialist and Object365 datasets for detecting objects in cluttered environments with an interactive frontend.
Get In Touch
I'm always open to discussing new projects, creative ideas, or opportunities to be part of your visions.
Let's Connect
Feel free to reach out if you're looking for a developer, have a question, or just want to connect.
Quick Contact
© 2026 Zaraar Malik. All rights reserved.