🔬 Semiconductor & Materials Science Background
PhD-level nanotechnology scientist with 10+ years of experience in statistical analysis of semiconductor manufacturing processes, material characterization, and device measurements. Expertise spans hypothesis testing, ANOVA, time-series analysis, and Six Sigma methodologies applied to tool development and process optimization.
📊 Data Science & ML Engineering
Specialized in tabular-data analysis and time-series modeling using the Python ecosystem (scikit-learn, XGBoost, PyTorch, PyMC3).
🔧 Data Science in Production
Experience building production ML systems with FastAPI, applying MLOps practices, and deploying scalable data solutions—from experimental design and statistical process control to predictive modeling and automated decision systems.
💻 Technical Problem Solver
Projects range from CUDA-accelerated Mandelbrot visualizations to WebSocket streaming dashboards and microservice-based prediction APIs. Advocate of clean code, thorough documentation, and user-centric design.
🔄 Continuous Learning & Experimentation
Data Science Learning Portfolio — Documented 5-year upskilling journey: 70+ projects, 50+ certifications (Imperial College, Google Cloud, AWS), progressing from Python foundations to production LLMs, and agentic AI systems, MLOps. Includes interactive D3.js project timeline. learning journey
🤖 Agents & LangGraph
Building stateful AI agents is my newest focus. I’m using the open-source langgraph framework to prototype multi-step, tool-using agents that persist state, support human-in-the-loop checkpoints, and recover from failures. My companion repo agent-lab (work-in-progress) collects reusable patterns—React-style planners, streaming memory nodes, and LangGraph + FastAPI micro-services—for anyone exploring agentic workflows.
🔍 Technical Interests
- Scientific Computing · NumPy, SciPy, mathematical modeling, algorithm optimization
- ML & Statistics · Experimental design, time series, Bayesian methods, ensemble models
- Visualization · Interactive plotting, real-time dashboards, GPU-accelerated graphics
- Web Development · FastAPI, WebSocket streaming, responsive data applications
🛠 Current Interests · Scientific computing • Algorithm optimization • Multi-angle dataset exploration • Async web APIs • Automated validation pipelines • Agentic systems
📖 Tech Stack Snapshot · Python • PyTorch • GBDT • scikit-learn • PyMC3 • NumPy/SciPy • Pandas • FastAPI • WebSocket • Docker • SQL • Langgraph • LangChain
📚 Past Exposure • MATLAB • R • C/C++ • Markdown • CSS • PHP • Assembly • Java
🥧 ML Modelling
Full evaluation of the adult salary dataset including EDA, modelling and evaluation. → dataset eval
Some raw model example based on the kaggle s5e12 competition dataset → raw models
🛎️ Services
FasrAPI microservice predicting restaurant tab (Docker-ready). → fast api1
FastAPI microservice predicting personality (Docker-ready). → fast api2
Streamlit rag agent, more complete service with tests and CI. → served chat agent
📊 Statistics
GLMM notebooks. → GLMM
🌀 Mandelbrot GPU
CUDA-powered fractal explorer with real-time zoom. → CUDA
📈 Streaming Plot
WebSocket server streaming data into interactive Bokeh charts. → Live plotting
