AI/ML Engineer • Turning data into production-grade intelligence
4+ years of experience building, deploying & optimizing machine learning systems

Core areas I work in:

• Deep Learning Frameworks → PyTorch (daily driver), Lightning, Transformers
• Large Language Models → fine-tuning (LoRA/QLoRA/DPO), RAG, inference optimization (vLLM, llama.cpp)
• Computer Vision → object detection, segmentation, OCR, multimodal models
• Data Handling → pandas, Polars, SQL, feature engineering
• MLOps & Deployment → experiment tracking, model serving, CI/CD for ML, Docker

Here are some repositories I'm particularly proud of:

• llm-fine-tuning-playbook
  Practical fine-tuning recipes (SFT, LoRA, QLoRA, DPO, ORPO) with modern best practices
  PyTorch • PEFT • bitsandbytes • Accelerate

• smart-document-processor
  End-to-end pipeline for invoice/receipt/form understanding (Donut, LayoutLMv3, Table Transformer)
  Transformers • PyTorch • OpenCV • OCR

• edge-vision-models
  Efficient vision models (MobileViT, EfficientViT, RepViT) optimized for mobile & edge inference
  PyTorch • ONNX • TensorRT • ncnn

→ Check out all my work → Repositories

Currently exploring / digging deeper into:

• Long-context & mixture-of-experts architectures
• Multimodal (vision + language) training & inference
• Extremely fast & memory-efficient inference (quantization, speculative decoding)
• Responsible AI, model evaluation, red-teaming

Last updated: February 2026