BUILD

CS @ UMBC · BS/MS

Hari Krishna
Amogh.

From computer vision to cloud-deployed apps, I work at the intersection of ML, systems design, and software that actually runs in production.

4 Projects built
1 Deployed on AWS
2 In progress
See the work
Python
React
AWS
PyTorch
OpenCV
SQL
Git
Linux
C++
Java
Pandas
MediaPipe

Certifications

Harvard University
CS50x — Introduction to Computer Science ↗ View credential
University of Alberta
Design Patterns ↗ View credential
Microsoft
SQL Foundations ↗ View credential
Pearson
System Design Fundamentals ↗ View credential
IBM
Introduction to DevOps ↗ View credential
Snowflake
Introduction to Generative AI ↗ View credential
PwC
Cybersecurity Consulting Simulation ↗ View credential

Projects

Live on AWS 2026

SortingΩ

Algorithm reference platform for sorting implementations across multiple languages. Built a community request system so developers can vote on which language they want next — making it a living reference, not a static page.

The interesting part: Deployed end-to-end on AWS. The request system collects real user input and feeds a backlog — treating a side project like a real product.
Built 2025

Stock Analyzer

Real-time market analysis app with live financial data, time-series analytics, and performance metrics — RSI, Stochastic RSI, volatility, moving averages. Not a toy dashboard; actual indicators traders use.

The interesting part: Chose Streamlit to ship fast without sacrificing interactivity — customizable inputs, dynamic charts, all from a clean Python stack.
Built 2026

Eyes — Facial Stress Tracker

Real-time computer vision system that detects stress by analyzing eye-openness, brow furrow, and jaw clenching. Outputs a live stress meter using sliding-window signal averaging for smooth, stable readings.

The interesting part: Fully modular architecture — detection, landmark extraction, tracking, and feature computation are independent components. Swap one without touching the rest.
Built · Team 2023

ASL Sign Language Recognition

5-person team project: real-time American Sign Language recognition converting hand gestures to letters. Trained on a Kaggle dataset with iterative preprocessing to improve accuracy across lighting conditions and hand shapes.

The interesting part: Accessibility-driven ML — the goal wasn't model accuracy for its own sake but making something that actually helps people communicate.

In Progress

In development 2026

Material Tracking System

Full-stack inventory system for an engineering firm — tracking materials in and out of warehouses in real time. Designed for actual operational use, not just as a demo.

  • Real-time material flow tracking across warehouse locations
  • Intake / dispatch logging with full audit trail
  • Dashboard for stock levels and movement history
  • Built for a real client in the engineering industry
Full-stackDatabaseReal client
In development 2026

Orbital Mechanics Simulator

An interactive 3D solar system simulator built with React and Three.js — pulling live orbital data from NASA's JPL Horizons API to render real planetary positions, near-earth asteroids, and Keplerian orbit paths.

  • Kepler's equations solved via Newton-Raphson iteration
  • Live ephemeris data from NASA JPL Horizons API
  • Near-earth asteroid tracking via NASA NeoWs API
  • Time controls — real-time to 1 year/second
ReactThree.jsNASA APIsVite

What I'm
building toward

I'm focused on building systems that hold up in production. The goal is simple: write software that gets used, scales, and solves real problems.

Work with me →

Ship more deployed projects

I focus on taking projects all the way to production, not just prototypes. SortingΩ is already live, and the Material Tracking System is being built with real operational constraints in mind.

🧠

Go deeper into ML/AI & Cloud

I'm focusing on applied ML and cloud systems — computer vision pipelines, model inference in real-time applications, and scalable backend deployment.

🏗️

Work on problems that matter

I want to build systems used by real users — tools that solve workflow, accessibility, or data problems in meaningful environments. The goal is impact through software that people actually rely on.