Hi, I'm Reyansh Joshi

Problem: DelveInsight needed a RAG-powered chat system that could answer questions from uploaded PDF reports and manage users, admins, and chat sessions at scale.

Solution: Built a React.js frontend and a Node.js/PostgreSQL backend for user management, admin controls, chat sessions, and PDF report uploads. Implemented the RAG engine in Python with FastAPI, LangChain, and OpenAI to serve contextual answers from report data.

Result: Reduced manual research effort by 60% and maintained 99.9% platform uptime.

Problem: Patient intake and hotel workflows were fragmented across modules with inconsistent data.

Solution: Implemented Node.js microservices across 3 integrated modules with transaction-safe database operations.

Result: Reduced inter-service API response time by 35% and improved data consistency.

Problem: UK healthcare revalidation platform needed a complete React frontend rebuild, but the client had no original Angular source code — only production build files and a live backend.

Solution: Reverse-engineered the UI and API behavior from production artifacts and GraphQL endpoints, then rebuilt the frontend in React with functional components and full feature parity. Fixed live-site bugs and optimized complex MongoDB aggregation pipelines.

Result: Delivered a working React frontend that lets doctors, admins, superadmins, and ROs manage revalidation workflows seamlessly.

Problem: Distributed service debugging relied on slow manual log investigation.

Solution: Engineered an intelligent log parsing and analysis engine using Python and FastAPI to automate root-cause detection.

Result: Reduced average debug time by 40% and saved 10+ engineering hours per week.

Problem: Conference rooms were double-booked and scheduling requests took too long.

Solution: Designed a full-stack scheduling platform with conflict-resolution logic on the MERN stack.

Result: Cut room scheduling time from 15 minutes to under 2 minutes per request.

Problem: Field teams needed a way to check sensor coverage areas and visualize them on a web portal instead of relying on manual verification.

Solution: Built a React web portal with JavaScript and Pyodide for in-browser sensor coverage analysis and visualization, backed by a Django/Python API with Celery for background processing.

Result: Enabled real-time sensor coverage checks and interactive visualization on the web portal, reducing manual verification effort.

Problem: IFAS needed a complete EdTech platform covering students, teachers, CRM, admin, test systems, live classes, and recorded lectures under one system.

Solution: Built a full EdTech ecosystem with student and teacher portals, CRM, admin panel, test and assessment systems, live classes, recorded lectures, real-time chat, and Razorpay payment integration.

Result: Handled 5,000+ concurrent users with zero downtime during peak live sessions.

Problem: Fragmented, manual onboarding process created delays and duplicated HR work.

Solution: Replaced it with an end-to-end MERN-based HR automation suite covering intake through completion.

Result: Streamlined onboarding for 200+ employees, improving process efficiency by 35% and cutting manual HR effort by 50%.