Selected Work

Building AI products backed by strong machine learning foundations.

My work spans AI product development and applied machine learning. From building end-to-end AI applications to developing predictive models, I focus on turning technical capabilities into practical solutions that solve meaningful problems.

Featured Product

Operational Intelligence Platform

🚨 Featured AI Product

Operational Intelligence Platform

Problem: Engineering teams often need to detect operational issues quickly and investigate incidents before customer impact grows.

I designed and built an end-to-end operational intelligence prototype that enables engineering teams to upload telemetry, detect anomalies, and receive AI-assisted investigation guidance. The solution combines a React frontend, FastAPI backend, Isolation Forest anomaly detection, and LLM-powered operational insights to demonstrate how machine learning can become a practical product.

Product takeaway: I designed the experience around operator workflows, trustworthy outputs, stable APIs, and explainable AI so engineers can confidently investigate operational incidents.

TypeScript React Python FastAPI OpenAI Isolation Forest Docker Vercel
Source Code Live Product

AI & Machine Learning Projects

These projects demonstrate my background in machine learning, deep learning, and applied AI. While they are technical implementations rather than end-user products, they shaped how I think about explainability, trustworthy AI, and building products powered by machine learning.

Skin Cancer Survival Prediction

🩺 AI / Healthcare

Multi-Modal Survival Prediction for Skin Cancer

Problem: Healthcare prediction is only useful when AI is understandable, not just accurate.

I developed survival prediction models using clinical, genomic, and imaging data, combining survival analysis, deep learning, and SHAP explainability to improve prediction transparency.

Technical takeaway: This project strengthened my understanding of multimodal learning, explainable AI, and the importance of building models that clinicians can understand and trust.

Survival Analysis Deep Learning SHAP Healthcare AI
Source Code
Emergency Vehicle Classification

🚑 Computer Vision

Emergency Vehicle Image Classification

Problem: Visual classification can help teams identify critical objects faster in high-volume environments.

I trained a convolutional neural network to distinguish emergency and non-emergency vehicles while evaluating model performance, accuracy, and generalization.

Technical takeaway: This project deepened my understanding of computer vision, CNN architectures, and deploying image classification models for practical decision support.

Computer Vision CNN Image Classification Model Evaluation
Source Code
Disaster Tweet Classification

🌍 NLP / Crisis Intelligence

Disaster Tweet Classification

Problem: During emergencies, critical information is often hidden within large volumes of noisy social media data.

I built an NLP classification pipeline using BiLSTM and GRU models to identify disaster-related tweets and improve signal detection from unstructured text.

Technical takeaway: This project expanded my understanding of NLP pipelines, sequence models, and using AI to surface meaningful information from unstructured data.

NLP BiLSTM GRU Text Classification
Source Code
3D contact envelope illustration

Contact me

If you’re interested in working together, don’t be a stranger.

Say hi!