Bimlesh

Software Infrastructure and Engineering

• Architected the backend for the Ocean Data Viewer (ODV) using Express.js and GeoServer, enabling offline deployments in secure, air-gapped environments.
• Designed a multi-tier deployment of 6 Docker Compose services behind an Nginx reverse proxy, packaged as a ~1.1 GB self-contained install for those air-gapped environments.
• Engineered high-performance GIS query pipelines using PostGIS ST_Intersects and GiST indexing, handling 50M+ rows per table and reducing query latency to milliseconds.
• Developed a fully integrated admin dashboard protected by secure JWT authentication, automatic dual-token rotation, and multi-tier RBAC.
• Built a state management system backed by Redis and MongoDB to persist complex user sessions and map viewports.

• Built a local-first Mac cleaner that finds regenerable clutter and large files, then confirms deletes before anything is removed. Paths stay on the machine.
• Desktop app is a Tauri and Rust scanner with a React UI. The marketing site and install flow run on Next.js with Clerk auth.
• Ships signed releases with curl and Homebrew install. Source is closed; public releases are the distribution path.

Tech: Tauri, Rust, React, Next.js, TypeScript, Clerk

• Next.js event portal built for traffic spikes from 5,000+ participants.
• 15+ reusable components and a SQL-backed flow for registrations and project submissions, with validation and data-integrity checks.

Tech: Next.js, TypeScript, PostgreSQL, Firebase, Vercel, AWS

• Built a consent-driven academic analytics platform in Next.js, with dashboards and peer comparison views for SGPA/CGPA trends.
• Implemented frontend-controlled visibility modes (Anonymous, Pseudonymous, Visible) with real-time UI updates, backed by Supabase Auth (OAuth).
• Enforced privacy boundaries with Row-Level Security and consent-gated APIs so grades are not public by default.

Tech: Next.js, React, Supabase, PostgreSQL, OAuth

• Architected a production-grade SaaS platform with a FastAPI backend, orchestrating data pipelines and third-party APIs for customizable retrieval workflows.
• Built an end-to-end search pipeline using FAISS vector indexing and Redis-based semantic caching, reducing retrieval latency by 60%.
• Implemented a spaCy-driven PII sanitization layer to filter sensitive data before processing.
• Designed a multi-tenant architecture using Auth0 for RBAC and MongoDB Atlas for knowledge-base management.
• Used Gemini and LangChain for customizable AI workflows and RAG on those tenant knowledge bases.

Tech: Next.js, FastAPI, MongoDB Atlas, Redis, FAISS, Auth0, LangChain, Gemini API

• Real-time market visualization using dynamic bubble charts, with size and color driven by performance.
• Covers multiple Indian indices with dataset switching, backed by MongoDB Atlas and Redis for low-latency delivery.
• Deployed on a VPS with Docker and CI/CD.

Tech: React, Express.js, MongoDB Atlas, Redis, Docker, WebSockets