Priyansh Singhal

I’m a Senior Full-Stack & AI Engineer focused on shipping production-grade systems that make LLM features dependable and affordable. At Couchbase, I cut LLM spend and response latency by building PipelineIQ, with multi-provider LLM abstraction, automatic model fallback, per-request token/cost tracking, and 24-hour response caching.

I also built Sentinel-AI, an ITSM platform that grounds ticket resolutions in RAG on Couchbase Vector Search with database-validated citations to reduce hallucinations, plus automated triage across 10 ticket domains. To make BI practical with LLMs, I created NLQDashAI, reducing LLM API spend by 60–80% via natural-language-to-SQL using Databricks server-side queries through MCP tool-calling.

Earlier, as a Full Stack Application Consultant (Contract) at Couchbase, I designed and shipped OrbiGraph, a multi-tenant workforce-orchestration SaaS delivering interactive org charts, headcount analytics, 9-box talent review, and compensation dashboards. I enabled non-technical analysts with an NLQ analytics engine where Claude executes SQL against Databricks via MCP, supported a provider-agnostic AI abstraction with quotas and usage tracking, and implemented dual-database persistence (PostgreSQL + Couchbase Capella) plus an Integration Studio with seven enterprise connectors.

Before that, at Oracle India Pvt Ltd, I built enterprise applications using Java, Spring Boot, Oracle APEX, React, and RESTful microservices, delivering layered APIs with tuned SQL for sub-second performance. I improved accessibility (WCAG 2.1 AA) and reduced deployment time by 30% through Jenkins-based CI/CD and API test suites, and I started in full-stack web development with React/Node.js/Express.js where I achieved sub-200ms API latency.