Welcome to my data portfolio! Here I document a summary of the projects I have contributed to.
I am more than happy to collaborate with you in projects. Please feel free to reach out 🤗
- experienced as a Machine Learning Engineer in research oriented & industrial level AI projects at Synopsys
- a first hire to a Global AI Research & Development Team at Synopsys
- proficient in Python
- avid self-learner
- love to code, break and then build from scratch
- authored a US patent and research publications in design automation conferences held in US & Taiwan
- BSc in Engineering (Hons) specialised Electronics & Electrical Engineering
- have expertise in Data Analytics, Data Engineering, Computer Vision, Deep Learning, Machine Learning & Large Language Models
- 💬 ask me about CI/CD pipelines, AI Integrations, data intensive applications & natural language processing with Transformers
- proud about my resillience I keep building up as a machine learning engineer starting from my final year computer vision research project 🦾
| Project Link | Video | Tools | Project Description |
|---|---|---|---|
| 🌍♻️Can I Bin It? Kaggle Write Up |
Gemini 3 Pro TypeScript |
This is an intelligent, agentic AI assistant designed to solve the confusion around recycling rules while inspiring a circular economy through creative upcycling. The spark for this project didn't come from a dataset or a policy paper, it came from a look on a face. One morning, watching the local waste collection, I noticed the bin collector pausing at our bin. She looked at the contents, shook her head with a look of sheer resignation and sadness, and had to tag the bin as "contaminated." It struck me: People want to do the right thing. Most contamination isn't malicious; it's confusion. We "wish-cycle", tossing things in the blue bin hoping they are recyclable, unknowingly ruining the whole batch. I realized that if we could bridge the gap between intent and knowledge and provide a creative outlet for the things that can't be recycled, we could reduce the burden on our waste workers, divert tons from landfills, and take a tangible step against climate change. |
|
| Gemini Powered Football Commentator⚽️ | Google-ADK Gemini pydantic |
This project implements a multi‑agent football commentary and research assistant AI Application. It demonstrates a complete agent‑tool orchestration pipeline, combining: - Data retrieval - Pydantic validation - Markdown report generation - Text-To-Speech audio creation - Multi‑agent coordination |
| Project Link | Achievements | Tools | Learning Log |
|---|---|---|---|
| Orchestrating Workflows for GenAI Applications | Airflow Weaviate FastEmbed |
Deeplearning.AI short course completed to deepen and refine my 4+ years of ML product delivery experience orchestrating ML prototypes with Airflow |
| Project Link | Web App Link | Tools | Project Description |
|---|---|---|---|
| Predict Accident Risk🚦🚗 | Streamlit App | Python Sklearn Pandas Streamlit Ensemble Learner Models |
Developed & deployed a machine learning model pipeline to predict the accident risk level (0-1) for a given set of past reported accidents, road infrasturcuture & environmental conditions. The regression model achieved a Root Mean Square Error (RMSE) of 0.05552 on the unseen test set, showing strong predictive accuracy. Since the road accident risk values range from 0 to 1, this low error indicates that the ML model’s predictions are very close to the true risk levels. Overall, the results demonstrate good generalization to new data. The stakeholders would be the transport authorities & self-driving car manufacturers |
| Project Link | Tools | Project Description |
|---|---|---|
| Travel Insurance Claims Analysis🪪⚖️ |  Python Pandas |
📊 Comprehensive exploratory data analysis on the Travel Insurance Dataset, focused on: - 💰 Claim acceptance patterns and customer behavior - 🧳 Product performance across agencies & agency types - - 💡 Market Segmentations |
| Project Link | Tools | Project Description |
|---|---|---|
| Claims Integrity and Risk Analysis🏥 | GCP BigQuery Looker Studio SQL |
This project focuses on detecting anomalies in hospital claim data to detect fraud and enhance payout efficiency Topics discussed: - Data Validation - Anomaly Detection - Fraud Logic Model |
| Project Link | Web App Link | Tools | Project Description |
|---|---|---|---|
| Personalized Anaesthesia Management🩺 | Streamlit App | Python Sklearn Pandas Streamlit Ensemble Learner Models |
Developed & deployed a machine learning model pipeline to estimate post-surgical outcomes for patients based on demographic, surgical, anesthetic, and perioperative data |