AI Software Engineer @ York Solutions · Client @ Optum
Building Multi-Model LLM Orchestration, large-scale AI systems, distributed infra & production-grade RAG Pipelines • MCP Servers • GenAI & Fine-Tuning • Vector Search • Cloud-Native Production
York SolutionsCurrent
ProPeers2 years 4 months
Cloud Conduction6 monthsI came from a Tier 3 college with no senior, no network and no roadmap. Just a hunger to figure it out.
I cracked remote SDE roles not because I was the smartest in the room, but because I built the right systems, followed the right patterns and never stopped shipping.
Today I build production Data and AI applications as AI Software Engineer @ York Solutions (Client @ Optum). Before that, as Founding Engineer at ProPeers, I architected agentic pipelines, RAG, MCP and multi-model orchestration in products that reach 600K+ users. Everything I learn, I teach, for free, to 40K+ engineers.
York SolutionsClient @ Optum · Indore, India · RemoteCurrentOngoingProPeersDelhi, India · Remote2 years 4 months2 rolesCloud ConductionUSA, · Remote6 monthsProduction conversational AI at princesinghai.com/arx · 24+ frontier models across 11 providers (GPT-5.4, Claude Opus 4.8, Gemini 3.1 Pro, Grok 4.3, DeepSeek V4, Kimi K2.6) · 3-Layer Hybrid RAG over 10M+ ChromaDB embeddings · ~99% RAG accuracy · sub-300ms first-token · zero-redeploy model registry
Multi-LLM orchestration arena at princesinghai.com/versus · one prompt raced across 2–24 frontier models concurrently · modelId-tagged SSE demultiplexing · live first-token latency benchmarking · 10M+ tokens benchmarked · shared 3-Layer RAG pass for fair-by-construction comparisons
Real-time voice conversations with 24+ frontier models at princesinghai.com/voice · full-duplex WebSocket pipeline (48kHz capture → 16kHz PCM → Azure STT → LLM → SSML TTS) · sub-200ms latency · 50+ languages · Voice Activity Detection · JWT via Sec-WebSocket-Protocol
Production AI platform · 3-phase architecture (Chat/Comparison/Voice) · 20+ LLMs (GPT-5.2, Gemini 3 Pro, Claude Opus 4.6, Grok 4.2, LLaMA 4 Maverick) · parallel 2-4 model comparison · voice-to-voice <300ms · RAG with 10M+ embeddings · MCP protocol · 50K+ daily API calls
Production Go AI DSA tutor · 6-step RAG (HyDE → 12 ChromaDB queries → RRF k=60 → MMR λ=0.7) · 10,000+ problems across 8 platforms · GPT-5.2 + Claude Sonnet 4.6 (Bedrock) + Gemini 3.1 Pro · SSE streaming · MongoDB AI cache · sub-200ms Fast Search
Production multi-cloud platform (princesinghdev.com + ai.princesinghdev.com) · AWS primary + Azure secondary HA · Next.js SSR/ISR · 98+ Lighthouse scores · hybrid DB (MongoDB + DynamoDB + Redis) · real-time notifications <50ms · 99.99% uptime · 18K+ AI queries processed
Preparation StreetEdTech & Mentorship Platform32K+ usersFull EdTech platform trusted by 40,000+ students · 1:1 personalized mentorship · structured roadmaps for DSA, development & system design · 150+ programming sheets · mock interviews · resume reviews · ₹1.3 Lakh+ revenue · 4.7/5 rating
AI systems at ProPeersShipped as Founding Engineer, serving 600K+ learnersSelf-learning RAG pipeline · ChromaDB vector search · text-embedding-ada-002 · ~99% accuracy · 600K+ users · <1s inference latency · personalized learning paths
60K+ usersAI code judge engine · Python/Java/C++/JS support · dual-layer JSON parsing · COMPILATION/RUNTIME/VALIDATION classification · deterministic verdicts · educational feedback
100K+ usersMCP-layered agentic assistant · O1/O3 intelligent routing · token metering system · 2x resolution speed · 3x engagement boost · context-aware responses
80K+ usersMonaco editor integration · <40ms real-time inference · inline reasoning display · multi-language execution · 40% retention boost · IDE-like experience
100K+ usersIntent-based discovery engine · 3-collection ChromaDB · GPT-5.2 + Kimi K2.5 · Hinglish query support · 19 smart chips · unified search across 5 content types
100K+ usersBy category
Layer 1 / 8
LLMsFrontier models routed per request: OpenAI, Anthropic, Google, xAI, Meta and more.Prince is a wonderful person. He listened to all my requirements very carefully without any interrupt. He then suggested a career option which was most relevant to my current situation, closest to reality. He is one gem and he is the best if you want either career or technical guidance
Great Session, It looked like he wanted to help me out and not just end the session after 30 mins. If you want a clear Idea on How to improve in DSA or How to become a better SWE then I highly recommend Prince.
I totally agree with this sheet; it's really very helpful. When I started DSA, problem-solving was tough for me, but after going through this sheet, my confidence grew, and my problem-solving skills improved significantly. Thank you so much, bhaiya!
This is a life-changing resource! I started DSA thanks to Prince, and I've been consistent for two months now. The guidance has been invaluable. I'm now looking forward to making a switch to development roles. Thank you for everything, bhaiya!
LeetCode1879+ (Top 5% Worldwide)1400+ solved
GeeksforGeeksInstitute Rank 1 & Global Rank 981300+ solved
InterviewBit1854+ (Master)560+ solved
Code3601854+ (Specialist)2000+ solved
HackerRank6-star Problem Solving300+ solved
HackerEarth1260+ Top 10%200+ solved