Sanjeev shankar
AI systems & SaaS builder, scaled platforms to 3M users
Ready for work
Sanjeev is ready for their next project!
🚀 AI SaaS Platform | Full-Stack Development
https://www.getconch.ai/
Overview
Built a scalable, multi-product AI platform combining writing, study, and content generation into a single ecosystem.
💼 What I Delivered
Developed a modular full-stack architecture using Next.js, Express, and MongoDB
Integrated multiple AI providers (OpenAI, Gemini, Claude, Azure) for optimized cost + performance
Built a rich AI writing editor with templates, rewriting, and structured content generation
Created an AI detection + humanization system with context-aware rewriting
Designed a study suite (flashcards, mind maps, planner) from PDFs, YouTube, and web content
Implemented an AI podcast generator with text-to-speech (MP3 output)
Developed a persistent AI chatbot across the platform
⚙️ Technical Strength
Scalable backend with rate limiting + autoscaling
Clean modular “mini-app” frontend architecture (fast feature shipping)
Strong data + file management system (folders, exports, storage)
Stripe billing integration with subscription lifecycle handling
Full analytics stack (PostHog, Mixpanel, Sentry)
1
125
🚀 No-Code Marketplace Builder | Full-Stack SaaS
https://marketplace.sanjeevdev.in/
Built a scalable no-code platform that enables users to create fully custom online marketplaces with a drag-and-drop builder, dynamic database, automation logic, email campaigns, and Stripe payments.
⚡ Key Highlights
Designed a visual canvas engine (drag-drop + responsive layouts) storing UI as JSON
Built a dynamic CMS where users create their own databases (auto-generated PostgreSQL tables)
Developed a no-code logic engine for filtering, conditions, and real-time data binding
Implemented multi-tenant architecture (schema-per-user) for strong isolation and scalability
Integrated email system (AWS SES) + Stripe subscriptions & webhooks
Added AI-powered UI generation using Anthropic Claude
0
71
🚀 Real-Time AI Voice Agent | Production System
Built a production-grade voice agent that conducts autonomous phone conversations using real-time STT, LLM, and TTS with sub-second latency.
⚡ Key Highlights
Developed a real-time voice pipeline (Twilio → Deepgram → OpenAI → AWS Polly → Twilio)
Enabled streaming conversations with instant transcription, reasoning, and speech synthesis
Implemented interruption handling (user speech stops AI mid-response seamlessly)
Built scalable load balancing across multiple API keys for concurrent call handling
Stored call history, transcripts, and AI summaries using DynamoDB
⚙️ Engineering Strength
Streaming-first architecture (WebSockets end-to-end) for ultra-low latency
Sentence-level TTS batching → natural, uninterrupted speech output
Direct audio byte streaming → minimal STT delay
In-memory session management → fast call state handling
Robust observability via AWS CloudWatch
🧠 Impact
Delivered a human-like conversational AI over phone calls
Achieved real-time interaction with interruption support (rare + complex)
Built for production scalability and reliability
0
52
🚀 Live Collaborative Coding Platform | Full-Stack
https://code.sanjeevdev.in/
Built a real-time coding platform where teachers manage classrooms, assign projects, and review student code live with a multi-language editor + compiler.
⚡ Key Highlights
Developed real-time collaboration using Socket.io (http://Socket.io) (room-based sync for each project)
Built a multi-language code editor (CodeMirror 6) supporting JS, Python, Java, C++, Go, PHP, Rust
Integrated live code execution via Judge0 API for instant output
Designed classroom + role system (teacher/student, join requests, project workflows)
Enabled multi-language persistence — users switch languages without losing code
⚙️ Engineering Strength
Room-based isolation → secure, real-time collaboration per project
Last-write-wins sync → optimized for speed over complex OT/CRDT
Prisma + PostgreSQL (Neon) → scalable, structured data layer
JWT + Zod validation → secure, type-safe APIs
Clean separation of REST (data) + WebSocket (realtime)
🧠 Impact
Delivered a Google Docs–like experience for coding
Enabled real-time teaching, pair programming, and code reviews
Built for speed, simplicity, and scalability
0
57
🚀 Real-Time Prediction Market | Full-Stack System
https://probo.sanjeevdev.in/
Built a real-time trading platform where users take YES/NO positions on live events (sports, crypto, politics) with instant order matching and live updates.
⚡ Key Highlights
Developed a microservices architecture (API Gateway, Trading Engine, WebSocket, Frontend) connected via Redis
Built a high-performance trading engine with in-memory orderbook and real-time order matching
Implemented complementary pricing logic (YES + NO = ₹10) with automatic liquidity handling
Enabled live orderbook updates using WebSockets + Redis Pub/Sub
Designed a responsive trading UI with portfolio tracking and one-click order execution
⚙️ Engineering Strength
Redis as message bus → decoupled, low-latency communication
Integer-based pricing (paise) → eliminates floating-point errors
Market maker bot → ensures liquidity with realistic price ladders
Optimized CI/CD → deploy only changed services via GitHub Actions
0
58