I build AI
products
end to end.
I'm Shubham Singh, an AI engineer. I design and write agentic systems, RAG pipelines, and data platforms myself, from the database to the UI.
Production ready.
Scalable backend.
No noise.
Four layers. One person building all of them.
Most people specialise in one. I work the whole stack — which is what it takes to get an AI product from idea to something a ministry or a business can actually rely on.
Frontend
- React & Next.js
- TypeScript / JavaScript
- Design systems
- Responsive UI
Backend
- FastAPI & Python
- Node.js services
- PostgreSQL
- APIs & integrations
AI Engineering
- RAG pipelines
- CrewAI / LangGraph
- Multi-agent orchestration
- Multilingual LLM systems
Systems & Data
- SQL, ETL & data pipelines
- Qdrant & vector search
- Context & token strategy
- Cross-session memory
Everything I ship with.
From crafting prompt injection defenses and orchestrating multi-agent architectures to building robust full-stack web and mobile apps.
Things I've built.
Data Intelligence Platform | QuadCentrix
An enterprise AI platform that turns raw organisational data into discoverable, quality-checked, and governed data that AI agents can safely use. Deployable on-premises or as SaaS, built on CrewAI, PostgreSQL, and multiple data connectors.
BolChat | Multi-Tenant RAG Chatbot SaaS
A solo-built B2B SaaS that lets businesses embed a multilingual, knowledge-grounded AI assistant on their website, with lead capture, analytics, and configurable personas.
Sakhi AI | Multi-Agent Safety Companion
A multi-agent AI companion for domestic violence survivors, offering legal information, safety planning, and NGO, shelter, and helpline lookup. Built with LangGraph and RAG.
AI Social Studio | AI Content Generation SaaS
An AI SaaS that turns trending news into platform-specific posts, memes, and blogs, with automatic fallback between LLM providers so generation keeps working during outages.
Where I've worked.
QuadCentrix
AI EngineerDesigned the architecture of the company's AI Data Intelligence Platform and drive its delivery across four modules: data discovery, data quality, process mining, and an AI context & governance layer. Built with CrewAI multi-agent workflows, Python, and SQL-based ETL pipelines. Also built Hindi/English voice and WhatsApp agents for government-sector use cases using Sarvam STT/TTS, Twilio, and LangGraph.
Rashi Technologies Pvt Ltd
Full Stack DeveloperBuilt and scaled Meri Bhakti, a faith-tech platform using Next.js, Node.js, TypeScript, and MongoDB. Developed secure REST APIs, payment workflows, and integrated analytics (Google Analytics, GTM, Meta Pixel). Started as a Frontend Developer Intern building responsive, SEO-optimized interfaces.