About Me

AI systems engineer & full-stack architect — building intelligent automation, RAG, and multi-agent systems

Abhishek Singh - AI Systems Engineer & Full-Stack Architect

AI Systems Engineer & Full-Stack Architect

I design and build intelligent automation systems that scale. With 7+ years in frontend architecture and performance engineering, I now specialise in applied AI systems — multi-agent orchestration, RAG, and async task pipelines.

I built Zaytri, a production-ready AI automation platform with dynamic LLM routing, brand-aware vector search (pgvector), and Celery-based background processing. I bridge product thinking, AI engineering, and scalable architecture — and previously led frontend teams delivering 90+ Lighthouse scores.

Core Expertise

AI & LLM Systems

  • • Multi-agent AI system design
  • • LLM orchestration & prompt engineering
  • • RAG & vector search (pgvector)
  • • Multi-LLM routing (Ollama, OpenAI, Gemini)
  • • Cost-optimized hybrid AI (Ollama + cloud)

Backend & Infrastructure

  • • FastAPI & async Python
  • • Celery + Redis task pipelines
  • • PostgreSQL + pgvector
  • • Docker, CI/CD, AWS
  • • Next.js, React, TypeScript

AI Systems & Observability

I ship production AI with clear metrics and reliability:

  • •Designing multi-agent workflows with observability and fallbacks
  • •RAG pipelines with brand-aware retrieval and tuned embeddings
  • •Async task queues (Celery/Redis) for scalable background processing
  • •Premium dashboards (e.g. Next.js) for AI workflows and cost/latency visibility
  • •Frontend performance: Core Web Vitals, 90+ Lighthouse, SSR/SSG

Leadership & Delivery

I lead teams and ship outcomes across AI and product:

  • •Leading cross-functional delivery of AI and full-stack products
  • •Mentoring on applied AI, prompt design, and production ML practices
  • •Architecting scalable systems that align with business and cost goals
  • •Working with product, design, and backend on roadmap and feasibility

DevOps & AI Infrastructure

I run reliable deployment and observability for AI and web apps:

  • •CI/CD with GitHub Actions, Docker, and AWS
  • •Deploying FastAPI, Celery workers, and vector DBs (PostgreSQL/pgvector)
  • •Monitoring, logging, and cost visibility for AI workloads
  • •Next.js, Vercel, and front-end performance (Lighthouse, Core Web Vitals)

How I Work

I choose tools and architecture to fit the problem: FastAPI and Celery for AI backends, Next.js or React for interfaces, and the right LLM (Ollama, OpenAI, Gemini) for cost and quality. I focus on product impact, observability, and maintainability rather than a single stack.