Fremont, CA · Open to opportunities

Saketh
Elumalai

High schooler building real things — AI apps, data pipelines, neural networks from scratch. Interested in electrical engineering and computer systems.

I'm a high school student at Irvington High School in Fremont, CA. I build software because I find the problems interesting — not because a class told me to. My projects span full-stack web apps, machine learning, and systems programming.

I'm drawn to the intersection of software and hardware: how systems are designed, how circuits compute, how data moves. I'm pursuing electrical engineering with a computer focus and I'm already doing the work.

01
I ship, not just code
AI tutoring apps, real-time wildfire maps — built end-to-end, deployed to production. Not just side projects sitting on a hard drive.
02
First principles thinking
Built a neural network in pure NumPy before touching any ML library. I want to understand how things work, not just make them work.
03
Broad, practical stack
Java, Python, C/C++, JavaScript, SQL, React, Flask, Docker. I pick up what a project needs and I'm comfortable across the full stack.
04
Early and driven
I'm in high school with production deployments, industry certifications, and college coursework already behind me. That momentum doesn't slow down.
Compify
Python · Flask · React · TypeScript · Vite · Docker
Sep 2025 – Feb 2026
AI-powered competition math tutoring web app — submit problems via text or image and get structured solutions. Integrated agentic RAG with thinking mode to retrieve AOPS problems, multimodal OCR for high-accuracy math notation transcription, and a PracticeModal component deployed to Vercel with localStorage-based API key persistence.
Wildfire Tracker
Next.js · NASA FIRMS API · Leaflet.js · PostgreSQL
Feb 2026
Real-time wildfire map for California using NASA FIRMS satellite data with 24-hour refresh. Implemented brightness filtering to surface high-intensity fire detections from live sensor data. Deployed to Vercel with environment-secured API keys and a public GitHub repository.
Neural Network from Scratch
Python · NumPy · Matplotlib
Mar – Apr 2026
Feedforward neural network in pure NumPy — no ML frameworks. Implemented forward pass, backpropagation, and gradient descent from scratch. Trained on 41,000 MNIST handwritten digit images achieving 89% accuracy over 500 iterations. Built ReLU and softmax activation functions manually.
Engineering Technology
YouScience Apr 2026 View ↗
3D Print Technology
YouScience Apr 2026 View ↗
Materials Engineering — Ohlone College
Ohlone College Aug 2025 Coursework
📄
Resume
Projects, skills, and education.
View →
✉️
Cover Letter
Tailored to the role I'm applying to.
View →
📬
Get in Touch
Reach out directly. I respond fast.
Contact Me →