Shi Sian Tan (Nick)

Quietly building intelligent systems so they can speak louder than I do.

Selected projects showcasing my work in industry, independent development, and research.

Professional Projects

University Curriculum Enabling Tool — Intel

Role: Developer (Collaboration)
Duration: Feb 2025 – Jul 2025
Repo: GitHub – University-Curriculum

  • Developed lightweight delivery platform with Next.js, FastAPI, and Payload CMS
  • Designed modular RAG pipeline for dynamic content delivery
  • Deployed multi-env Docker setup with GitHub Actions
  • Optimized for Intel XPU devices (edge inference)

Enterprise GenAI Platform - Intel OPEA

Role: Developer (under Tech Lead Direction)
Duration: Nov 2024 – Jan 2025
Repo: GitHub – GenAIExamples

  • Optimized containerized GenAI inferencing on Intel iGPU/dGPU
  • Migrated workflows from systemd → Docker Compose
  • Reduced setup friction, enabling faster adoption across teams
  • Delivered internal PoC demos to showcase inferencing capabilities

Workload Services Framework - Intel WSF

Role: Validation Engineer
Duration: Jul 2024 – Oct 2025
Repo: GitHub – Workload Services Framework

  • Supported validation of AI workloads on Intel Xeon platforms
  • Designed and executed test cases, triaged bugs with JIRA
  • Helped improve platform stability and coverage of edge cases

Independent & Personal Projects

End-to-End Regression MLOps Pipeline

Role: ML Engineer
Duration: Jul 2025 — Present
Repo: GitHub – mlops
Tech: Docker, Kubernetes, Machine Learning, MLflow, Github Actions, Python, FastAPI, RESTful API, scikit-learn, Prometheus, Grafana, Minio, Postgres, DVC, more

  • Full ML lifecycle pipeline for regression tasks: data processing, training, CI/CD, deployment, monitoring.
  • Experiment tracking with MLflow, model/data versioning with DVC, and automated retraining workflow.
  • Deployed as containerized services (Docker, Kubernetes); integrated observability (Prometheus / Grafana).
  • Result: measurable deployment lifecycle improvements and reproducible experiments for production-like testing.

AI-Powered Password Manager

Repo: GitHub Tech: Python, scikit-learn, Cryptography, SQL (hashed + salted storage)

  • Integrated AI-based strength evaluation into a desktop password manager
  • Used XGBoost + TF-IDF to assess and suggest password improvements
  • Built with Tkinter GUI for user interaction

Cost-of-Living ML Dashboard

Repo: GitHub Tech: Python (ETL), Tableau

  • Designed ETL pipeline with data cleaning and feature engineering
  • Built forecasting models for global cost-of-living trends
  • Published interactive Tableau dashboard for visualization and insights

Academic & Research Projects

Final Year Project – Colorizing Grayscale Images with Deep Learning

Role: Researcher
Duration: Oct 2022 – Jul 2023
Tech: Deep Learning, Keras, Tensorflow, c-GAN, CNN, Auto Encoder, U-Net

  • Proposed and implemented “Colour Stack cGAN” architecture
  • Trained and evaluated model on paired grayscale ↔ color datasets
  • Explored applications in image restoration and media colorization
  • Presented findings at academic symposium

Bank Building Classification System

Role: ML Engineer Tech: ML pipeline, OpenCV, PCA, SVM, scikit-learn, Numpy, Pandas, Seaborn, more

  • Led a team to build an image classification system for identifying bank buildings from online images
  • Applied PCA for dimensionality reduction and trained SVM classifier

Pattern Recognition – Multi-Label Classification

Role: ML Engineer Tech: ML pipeline, OpenCV, PCA, SVM, scikit-learn, numpy, pandas, seaborn sns, more

  • Preprocessed large facial image dataset
  • Applied PCA-based dimensionality reduction
  • Built ML pipelines for gender, emotion, age, and nationality classification

Notes

  • Each entry above links to the most relevant public repo when available. For internal/closed-source work I can provide summaries, architecture diagrams, or code excerpts on request.
  • If you want compact, CV-style one-line summaries of these projects (for LinkedIn or an emailed portfolio), tell me which items to shorten and I’ll produce a condensed version.

Last updated: Sep 2025