dharshan bellan

ai fullstack software engineer, hackathon winner, curr swe @ motive

my projects

some cool stuff i've built

yaptome

work in progress

reflect. improve. speak it into existence. a nightly voice AI that helps you process your day, plan tomorrow, and become the person you want to be — one conversation at a time.

Next.jsTailwindSupabaseVapiGemini

tiktok clone

built a full-featured tiktok clone with ai-powered video recommendations. uses vector embeddings to suggest videos you'll actually want to watch.

Next.jsShadcnTailwindLucidDrizzleNeonVercel BlobOpenAI

View Code

vzn - pinterest with ai

what's your [vzn]? it's like pinterest, but for ai-generated images. create and discover cool ai art in a beautiful social platform. i even hosted the image generation model myself using modal for serverless compute.

Next.jsTailwindFirebaseHuggingFaceSDXL-TurboModal

View Code

Live Demo

kupl - ai relationship analysis

ever wondered what your instagram dms say about your relationships? kupl analyzes your messages and gives you the full scoop - from most used words to relationship scores and summaries.

Next.jsTailwindFirebase

View Code

Live Demo

reviewpal - video & product summary scraper

a web app that scrapes youtube videos and product pages to get concise summaries. search for products or videos, and get a summarized version of the content, making it easier to gather key information quickly.

ReactTailwindNode.jsExpressPuppeteerCheerio

View Code

Live Demo

resume

Dharshan Bellan

SF/Bay Area | dbellan1291@gmail.com | linkedIn.com/in/dharshanbellann | github.com/virocodes | dharshan.dev

Experience

Software Engineer, QA

Mar 2025 - Present

Motive • San Francisco, CA

Built a computer vision pipeline for test video analysis, cutting manual review time by ~30% across 100+ tests/week.
Added support for a new vehicle simulation protocol, expanding coverage from 2 to 3 simulation types
Rebuilding reporting tool into a full-stack FastAPI web app for 20+ engineers, enabling interactive test result reporting
Collaborate with QA, embedded, and AI teams to build scalable automation tools and improve test system reliability.

Software Engineering Resident

Jul 2024 - Present

Headstarter • San Francisco, CA

Building 14+ machine learning, ai-engineering, and full-stack projects in fast-paced software team environments
Developing neural networks in Python, and 11 apps in Typescript on AWS/Vercel with dev and production environments
Implementing agentic AI, llm-chaining, etc, on 10+ LLM models
Coached by Google Machine Learning, Google Kubernetes, Two Sigma, Tesla, Figma, and Citadel Engineers

Software Engineering Intern

Jun 2024 - Aug 2024

MandrakeTech • Remote

Developed a full-stack web app using Flask and PostgreSQL to help users manage job applications
Implemented ScrapingDog API to scrape job listings from LinkedIn, enhancing job search automation
Deployed the application on Heroku, achieving 50+ visits and 30+ user sign-ups

Machine Learning Intern

Nov 2023 - May 2024

S#AI • Remote

Built a GPT model from scratch in PyTorch, implementing tokenization, training, and inference
Trained a character-level GPT model using AdamW optimizer on a T4 GPU in Google Colab to generate text

Software Engineering Intern

Mar 2024 - May 2024

Endgrate • Remote

Developed a Flask-based API integration tool to automate data transfer between Asana and Todoist

Projects

storybook

Oct 2024

Team Project (~32 hours)

Won $1000 (2nd Place Best AI App) from Story Foundation for an interactive AI-powered document builder
Designed a canvas-style UI in Next.js and Tailwind with draggable blocks and embedded chatbots
Integrated OpenAI API to return JSON-formatted data, dynamically rendering tables in HTML

TikTok Clone

Dec 2024

Team Project (~30 hours)

Built a TikTok-style recommendation system using vector embeddings for personalized video feeds
Developed a scalable recommendation pipeline with Next.js, TypeScript, Neon, Drizzle, and OpenAI embeddings
Optimized user engagement tracking using cosine similarity and a learning rate-based update strategy

vzn

Aug 2024

Published Web-App 20 Users (~30 hours)

Created an AI-powered vision board generator with Next.js and Firebase, reaching 20+ users
Implemented Stable Diffusion sdxl-turbo model on HuggingFace via Modal with sub-30s latency
Optimized cloud storage using Cloudflare R2 and AWS S3 API for efficient image retrieval

ReviewPal

Dec 2024

Team Project (~30 hours)

Developed a full-stack application with React, Node, and Express to generates product summaries from YouTube videos
Scraped video transcripts using YouTube Data API and generated summaries with Gemini API

aptrak

Jun 2024 - Aug 2024

Published Web-App 50 Users (~40 hours)

Developed full stack job tracking application using Flask, PostgreSQL, and Jinja
Implemented OpenAI API to generate AI-powered cover letters from job descriptions and resumes
Designed an intuitive UI with JavaScript, HTML, and CSS, enhancing user experience

Skills

Languages/Frameworks

Python (4yrs) · JavaScript (3yrs) · React (2yrs) · Next.js (1yr) · Node.js (1yr) · Flask (1yr)

Software

Firebase · Vercel · Heroku · JupyterNotebook · Google Colab · Git · Bash

Education

Diablo Valley College

Expected Graduation: May 2026

Associate of Science in Computer Science | 4.0 GPA

let's chat

always down to talk about new projects, cool ideas, or just chat about tech. hit me up!

dbellan1291@gmail.com