Work
A mix of client work, internships, and things I built because the problem would not leave me alone — commerce flows, automation, ML-adjacent pipelines, and the occasional desktop experiment.
I care about correctness at boundaries (schemas, webhooks, idempotent jobs) and about systems that still make sense six months later.
Always happy to chat about new ideas — hit me up at@zendrix396.
SocialKart · AI-Powered E-commerce Automation
End-to-end pipeline that scrapes Instagram content, classifies imagery, analyzes transcripts, and generates marketplace-ready listings with Gemini — built for the Amazon Sambhav Hackathon ecosystem.
- Multi-modal ingestion: video, audio, and images from social posts with structured product output for sellers.
- Flask + Socket.IO backend with Docker packaging; ONNX Runtime for image-side workloads alongside LLM calls.
- Real-time progress over websockets; EC2 deployment when workloads need sustained throughput.
- Selected among the Top 5 teams in the Amazon Sambhav Hackathon.
view on github →test live →
VideoVision · Video Enhancement Pipeline
Quality-focused upscaling that leans on Real-ESRGAN, optical-flow propagation, and selective frame recomputation so GPUs spend time only where pixels actually change.
- Propagates enhancements across temporally stable regions instead of blindly re-rendering every frame.
- Optional GFPGAN restoration path for portrait-heavy footage that benefits from face-aware cleanup.
- Tuned for long clips where naive frame-by-frame super-resolution would be prohibitively expensive.
view on github →
Health-Vault · AI Medical Analysis System
An advanced system that analyzes medical reports, predicts potential ailments, and suggests medicine using custom ML models.
- Analyzes uploaded medical reports (PDF/Image) using PyPDF2 and OCR, with an LLM for summarization.
- Features a dual-model system for disease identification and medicine recommendation based on user-provided symptoms.
- Built with a FastAPI backend allowing for dynamic model retraining and loading via Excel data uploads.
view on github →watch demo video →
Fluent · Data analysis and function generation
Built a tool to turn raw data into accurate mathematical functions with a clean, interactive interface.
- Integrated CSV, XLSX, and JSON support with both file upload and manual input options.
- Implemented automatic regression-based function generation and interactive visualizations.
- Added LaTeX based pdf export and prediction capabilities for practical data analysis workflows.
try here →
frontPage · Paper Trading Platform
Django brokerage simulator with scheduled jobs and Redis snapshots so concurrent updates still yield accurate realized and unrealized P&L.
- Celery workers keep portfolios, fills, and pricing models in sync without blocking the web tier.
- Redis-backed caching for live quotes with guardrails around racey writes during volatile windows.
- GitHub Actions for CI so refactors stay shippable as the trading surface grows.
view on github →
Nyxx · Desktop Workflow Automation
Tauri + Svelte control plane over Rust Axum services that records and replays keyboard, mouse, clipboard, and window-level automations with isolated execution sandboxes.
- Macro library with placeholder-driven inputs so non-developers can reuse flows safely.
- MCP-compatible surface so agent clients like OpenClaw can discover and trigger automations without bespoke integrations per app.
- Designed for operators who need reliable desktop glue instead of brittle one-off scripts.
view on github →
Utilities and bots
A collection of specialized tools, engines, and bots built to solve specific problems or explore unique technologies.
- Symbolic Math Utility: A web app for symbolic math operations with a real-time UI built on a modern frontend stack.
- Krona-lite: A blazingly fast, minimalist desktop To-Do app built with Svelte and Tauri. (View on Gumroad)
- ConvolutionalChess: A playable chess engine in Python combining classic search algorithms with a modern neural network.
- Stealth Crawler: A traffic emulator using Puppeteer and residential proxies to mimic human browsing behavior.
View these and more on my GitHub.
B.Tech, Computer Science
KIET Group of Institutions, Ghaziabad
Sep 2024 – Sep 2028 (expected) · CGPA 8.58 / 10.0
Coursework highlights: DBMS, Computer Networks, advanced data structures — plus ongoing core CS per curriculum.
Certifications
- Pearson MePro Level 10 Expert — Apr 2025
- Cisco Networking Essentials — Mar 2026