An aspiring AI/ML Data Analyst & Full Stack Developer passionate about building sleek web experiences
I'm currently pursuing my B.Eng. (Honours) in Computer Science with a specialization in AI & Data Science at Nanyang Technological University (NTU), Singapore. My journey in technology began during my International Baccalaureate Diploma Programme at NPS International School, Singapore, where I developed a strong foundation in analytical thinking and problem-solving.
I'm passionate about leveraging technology to solve complex real-world problems. My interests span across Machine Learning, Computer Vision, Data Analytics, Full-Stack Development, and Climate Risk Assessment. I aspire to build scalable AI-driven applications that can contribute to sustainable development and create meaningful change in society. Through my internships and projects, I've gained hands-on experience in deploying production-ready solutions that optimize performance and deliver measurable impact.
When I'm not coding or training models, you'll find me on a quest for the perfect South Indian meal! There's something absolutely soul-healing about a steaming bowl of sambar with fluffy idlis, or the satisfying crunch of a perfectly golden dosa paired with coconut chutney. Hot rasam is basically my comfort code debugger, and crispy vadas are my favorite SNACK! I'm always down for exploring authentic Tamil Brahmin-style vegetarian cuisine — if you're up for some good food and even better conversations about tech over traditional South Indian flavors, count me in! 🍛✨
Professional experiences that shaped my career
Climate Risk Web App — Data & Dashboards
Ingested multi-hazard REST feeds into Postgres and exposed key metrics via Rails APIs for a React dashboard.
Deployed the app on Azure with Docker, optimising asset bundling and code-splitting to cut Largest Contentful Paint (LCP) from 2.8s to 1.7s.
Hardened system monitoring and consolidated logs, reducing incident-triage time by 35% across services.
Software Development Intern — Automation
Built a Java tool to auto-generate test-case matrices from requirements, cutting QA preparation time by 60%.
Shipped an ID-mapping service with Excel (Apache POI) export, eliminating manual matching errors and reducing support tickets by 30%.
Delivered a Swing CRUD UI with quick inline guides that onboarded new users in under one hour per session.
Process Automation (Excel/VBA)
Automated a Forms-to-Excel ETL pipeline, processing 70+ records in under 10 seconds with full audit trails.
Built a tool to auto-generate 36 letters and batch-create PDFs in under 3 minutes, saving HR 2+ hours per week.
Standardised the payment workflow, reducing claims processing time from 30 minutes to just 5 minutes.
Research Analyst — Singapore Food Study (Alt-Protein)
Analysed a nationwide survey (N=3,148) for the SFA's "30 by 30" initiative; found 60% higher willingness for plant-based vs insect-based protein.
Ran a messaging test (N=811); framing alt-protein as "tastes like meat" lifted purchase intent by 22% (p<0.05).
Tracked social discourse across 100+ campaigns, identifying a 3x growth in alt-protein conversation from 2022 to 2023.
Deep Learning Engineer (Computer Vision/OCR)
Deployed YOLOv8-based defect and OCR models, cutting production line inspection time by 50%.
Achieved 99.8% character-level accuracy (0.2% CER) on a proprietary test set of 500 printed-label images.
Automated the end-to-end ML pipeline (ingest, label, train, evaluate), reducing model iteration time by 40%.
Developed a custom PyQt annotation tool that increased labelling throughput and consistency for the team.
Javascript
React
Tailwind CSS
HTML
CSS
Framer Motion
C
C++
Bootstrap
Python
Java
Streamlit
Node.js
Next.js
Express.js
React Router
Nodemon
Ruby on Rails
Flask
FastAPI
Mongoose
MongoDB
Oracle
PostgreSQL
MySQL
Firebase
Git
Docker
GitHub
Google Cloud
AWS
Microsoft Azure
Jupyter
Postman
Scikit-Learn
PyTorch
Pandas
NumPy
OpenCVProfessional certifications and continuous learning achievements
Amazon Web Services (AWS)
Oracle Corporation
Imperial College London
DeepLearning.AI
CU Boulder / Stanford
University of Michigan
INSEAD
Udemy
Building communities, leading teams, and creating impact
NTU SCSE TOP
Led 40 facilitators for a 5-day "Camp Interstellar" (600 freshmen; S$12k); zero safety incidents and +18% post-camp engagement.
NTU Welcome Centre
Delivered 40+ sustainability-themed VIP tours; 9 corporate leads; trained 12 juniors (25% prep time).
Economics & Business Club (NPS)
Built a 200-member club; monthly FinTech workshops (92% turnout); secured S$2k sponsorship for a first inter-school case competition.
LCC
Served as an academic mentor for NTU students to improve their written assignments.
My latest contributions and coding activity.
Academic & Selected Projects
Aug 2025 – Jun 2026 | Biomedical Informatics Lab, NTU
Integrating AlphaFold/ESM-2 features with GNN/Transformer for Antibody–Antigen binding affinity prediction.
Perception + Control Role: Data & Model; Embedded Integration
Trained a custom YOLOv5 model and deployed real-time inference on RPi/OpenCV for object-based maze navigation. Integrated STM32 motor control and Android UI.
OpenCV
Frontend Development Role: UI/UX, Feature Modules
Built user modules (auth/profile), remittance calculator with live FX rates, and curated news feed. Integrated MITRA chatbot and multilingual support.
React
REST APIs
i18n
Console Application Role: Application Coder; Documentation Lead
Implemented a modular MVC CLI for camp creation, registration, approvals, feedback, and reporting with clear service boundaries.
Java
MVC
CLI
R²=0.83
Trained on 26 seasons; –20% MAE vs baseline; simulated +8% target-day lift in historical what-if tests.
+15% vs public projections
Back-tested 3 seasons; +15% over public projections on hold-out weeks; shipped explainer charts for lineup decisions.
Minority F1 +15%
Improved minority-class performance +15% via SMOTETomek resampling and model tuning on 6,497 samples.
Regression (MAE/MAPE)
Scraped 5,000+ postings across 50 cities; engineered text+structured features; reduced MAE vs baseline.
+10% over baseline
Compared models on 700+ records; +10% diagnostic accuracy over baseline; streamlined Spark preprocessing.
Forecast error (MAE/MAPE)
Modeled 500k+ transactions; beat naive seasonal baselines on MAE/MAPE; delivered store/category forecasts for inventory planning.
Some exciting ideas I'm currently working on.
A modern social media platform featuring user profiles, posts, real-time chat, notifications, and interactive feeds built with the MERN stack and Next.js.
Coming SoonA MERN-based admin dashboard to manage students, attendance, teachers, and performance data with charts and analytics.
Coming SoonA property rental platform with user authentication, property listings, filtering, booking system, and admin control panel.
Coming SoonLet's discuss your next project or just say hello!
I'm always open to discussing new opportunities, interesting projects, or just having a chat about technology and development.
AISH0021@E.NTU.EDU.SG
Location
Nanyang Technological University, Singapore, Singapore
Response Time
Within 24 hours