Mayank Manjeara

I am a senior software engineer who designs and ships production-grade AI systems and scalable backend platforms. I blend research and engineering to turn ML prototypes into reliable, cost-efficient production services.

At my current role I led development of a production AI/RAG platform using LLMs, vector search, and domain-tuned embeddings, achieving 90%+ extraction accuracy and a domain-optimized retrieval system with 98%+ top-k precision. I built semantic chunking and tagging frameworks that enabled lossless extraction from 97%+ of sites and reduced manual operations by 80%.

Previously, I built a SaaS foreign-exchange trading platform and broker module used across multiple applications, enabling generation of 10M+ rates daily and pioneering monitoring that analyzed 18M+ events per day. I also revamped a B2B onboarding flow, streamlining onboarding for 14 teams and architecting a flexible state machine to support varied organizational roles.

I lead cross-functional teams, designed event-driven orchestration and a horizontally scalable Celery mesh, and integrated observability with OpenTelemetry and SigNoz to cut inference costs and deployment time. I am passionate about building ML-first workflows, improving reliability, and delivering measurable business impact.