
Hi, my name is
I’m a full-stack AI engineer specializing in building highly accessible, human-centered applications for the web. Currently, I'm working at Onetera as an AI Engineer.
I'm Meghana — a Full Stack AI Engineer at Onetera who builds production AI systems that governments depend on daily.
Most engineers inherit systems. I built mine from scratch — as one of the founding engineers who took a government AI platform from zero to production. I own the full stack: agentic LLM orchestration, RAG infrastructure, AI observability, and fairness evaluation. The systems I've shipped have reduced citizen task resolution from days to under 15 minutes, cut costs by 60%, and achieved 50% fewer hallucinations in production.
Before AI, I architected enterprise platforms at Accenture serving 3,000+ users and built real-time IoT systems at Bosch — which is why I think in systems, not just features.
I graduated top 1% of my class at Cal State Fullerton with a 3.97 GPA, was recognized as best paper presenter at a national conference for ML research, and outside of engineering I find balance through dancing and painting.


August 2024 - Present
Most AI demos don't survive contact with government infrastructure. I joined Onetera as a founding engineer to make sure ours did.
I'm building the AI backbone for government-scale citizen services — systems that real people depend on to navigate housing, benefits, and public records. The stakes are different here. Latency isn't a UX metric; it's the difference between a citizen getting help or giving up.
I architected a semantic search engine across millions of government documents — vector-indexed, retrieval-benchmarked, returning accurate answers in under one second at scale. I built a multimodal AI compliance layer with model-tiering and confidence-based escalation that cut LLM costs by 60% while keeping SLAs intact. And I institutionalized a cross-model reliability framework — hallucination detection, response-consistency scoring — that made our outputs trustworthy enough for government deployment.
I engineered end-to-end traceability across every AI decision — full observability from prompt to response, so every output could be audited, explained, and trusted. Reliability wasn't a feature; it was the foundation — built into the infrastructure so the system held under real-world load without human babysitting.
What I care about most here is the constraint: build AI that's fast, honest, and explainable — not just to engineers, but to citizens and policymakers who have every reason to distrust it.


January 2023 - May 2024
Instructed 150+ students in learning React, covering key topics like component-based architecture, API integration, and state management, while guiding them in utilizing Git and GitHub for version control and collaboration.


August 2020 - August 2022
Before AI dominated engineering, I learned what it means to build systems that actually scale.
At Accenture, I operated above my title — making architecture decisions, defining engineering standards, and presenting technical roadmaps directly to enterprise client leadership.
I architected a mission-critical platform consolidating 30 subsidiary applications into one unified system — serving 3,000+ daily active users on Azure with zero downtime. I built a React Component Library adopted as the org-wide engineering standard, eliminating 40% UI fragmentation across a large-scale POS platform. I drove architectural decision-making across engineering and client leadership, translating complex technical tradeoffs into executive-level roadmaps.
The most important thing I learned here wasn't technical. It was that great engineering is about decisions that hold up at scale, under pressure, with real business consequences.
Stack: React · GraphQL · Express.js · Azure · MySQL · PostgreSQL · Kubernetes · Terraform · Splunk · Jest


June 2018 - July 2019
Designed and integrated 70+ RESTful APIs with Express.js — establishing API contracts and documentation standards via Postman and Apiary that became the engineering baseline for frontend-backend communication.
Engineered a real-time industrial IoT dashboard on AWS serving 1,500+ users — reducing rendering complexity by 30% and improving user interaction by 25% using Angular.
Architected and implemented enterprise-grade RBAC with JWT authentication — eliminating unauthorized access incidents and establishing security standards across production systems handling sensitive industrial data.
Next.js
GCP
Prisma ORM
Open AI
Stripe
ShadCN
Tap on a project to learn more!






React is my preferred library for crafting user interfaces. My proficiency extends to React Hooks, State Management Techniques, Context, and Redux, which I've applied both in my professional endeavors and in my current role as a graduate teaching associate.


2022 - 2024
GPA: 3.97
If you would like to work together or discuss an opportunity for work, please use the form or send me an email on meghanabtn@gmail.com