hieee,I'm Deepak
about
tl;dr: learning things by fiddling around the internet!
I dive deep into technologies, whenever something sparks my curiosity, I go all-in, often losing track of everything else around me!
Beyond code, I have a keen eye for the artistic side of tech. UI/UX, Cinematography, Photography, and Video-editing fascinate me, and I love blending creativity with technical precision.
technical skills
design skills
projects
Ecom-microservices
Developed a high-performance e-commerce microservices system handling 150+ requests/second with sub-320ms latency using Node.js, Redis caching, and optimized service design. Implemented robust distributed transactions with the Saga Pattern and Circuit Breakers, ensuring consistency across Order, Product, and Payment services. Secured the platform with Stripe idempotent payments, JWT-based RBAC, and IP rate limiting.
Distributed-task-scheduler
Built a high-performance distributed task scheduler in Go with a bounded worker pool capable of executing 50+ parallel jobs while maintaining stable memory usage. Designed a fault-tolerant distributed queue using atomic Redis operations and a recovery “Reaper” process to guarantee at-least-once delivery even during worker failures. Added a distributed rate limiter powered by Redis Lua scripts, ensuring atomic token-bucket enforcement and protecting downstream services from traffic spikes.
Agent-Hub Platform
A full-stack AI chatbot system that orchestrates multiple specialized agents, each driven by its own custom system prompt and isolated conversation history. It leverages a modern Next.js and React frontend with a TypeScript Express backend, integrates PostgreSQL via Prisma for robust data modeling, and implements JWT authentication with securely hashed credentials using bcrypt to deliver production-ready, project-based chat experiences powered by the Perplexity AI API.
File-Sharing Service
A Go-based service that lets users upload files (PDFs, images, ZIP archives up to 100 MB) via a web interface and stores them in Cloudflare R2 cloud storage. Each upload generates a unique download link backed by pre-signed URLs that automatically expire after 15 minutes, ensuring secure, time-limited access.
Screen Recorder
A privacy-first, fully client-side screen recorder built with Next.js & TypeScript that captures screen, webcam overlay, and multi-source audio entirely in the browser, with zero data sent to any server. Delivers professional-grade controls (custom FPS, bitrate, webcam positioning, pause/resume) powered by the native MediaRecorder API, with instant MP4 download.
experience
RL Platform Engineering Intern
WootzApp
Trained, Tested and Optimized RL workflows and improved agent reward functions to enhance performance and reliability. Debugged and improved core RL pipeline and benchmarks for the training environment. Worked with Tinker to train on various tasks with SWE-Bench and D2C to achieve state-of-the-art performance.
AI-ML Intern
Edunet Foundations
Completed an AI/ML internship at Edunet Foundation, where I explored core machine learning libraries like Scikit-learn, Pandas, and Keras. Built and fine-tuned models to achieve accurate predictions. Developed a leaf disease detection system that identifies plant diseases from images.
writing
I've begun writing blogs to share what I learn along the way, more exciting content coming soon!
connect
hellodeepakch@gmail.com