I work primarily with Python, backend systems, and applied machine learning.
My experience spans model development, deployment, and performance-aware workflows in production settings.

• Backend services using Python
• ML model integration and inference pipelines
• Automation and workflow orchestration
• Improving reliability and performance of existing systems

• Advanced Python design patterns
• CUDA fundamentals and kernel-level thinking
• GPU-based image processing (CuPy)
• Performance optimization using C++, Cython, and Numba
• Efficient ML inference and system-level tradeoffs

• Python-based tooling or libraries
• ML inference or deployment-focused projects
• Automation-heavy systems or data workflows
• Performance and optimization-oriented work

• CUDA and GPU performance debugging
• Low-level optimization strategies
• Python performance improvement with C++/Rust

• Python backend development
• Image processing
• Async task systems
• ML model training, fine-tuning, and deployment
• Data pipelines for image-based ML systems

• Frameworks: Flask, FastAPI
• Databases: MongoDB, SQLite
• Async / Queues: Celery, Redis
• Cloud (GCP):
  • CloudRun
• Cloud (AWS):
  • EC2
  • Lambda
  • ECR
  • SageMaker (training & serverless inference)
• Containers: Docker (image size & runtime optimization)

• GitHub issues or discussions on relevant repositories

• Python + Computer Vision
• Pandas, NumPy, OpenCV
• PyTorch (model creation, training, fine-tuning, debugging, custom loss, optimizations)
• ONNX conversion and inference workflows
• Image data collection, labeling, and cleaning
• Automation and complex workflows
• Performance optimization (time & memory)
• CUDA & GPU programming concepts
• TorchScript / TensorRT
• Triton inference
• C++ performance optimization
• Python acceleration with Cython and Numba

• Some production ML work cannot be shared publicly due to confidentiality.
• My work reflects active learning and hands-on experimentation rather than formal specialization.