Yash Varma is an AI Product Engineer who builds full-stack products with AI at the core, taking production systems end to end from first prototype to deployment, with 10+ national hackathon wins, peer-reviewed research publications, and experience engineering platforms at Norxio, ConnectWise, Cartman Labs, and EaseworkAI. He is currently based in Dublin, Ireland pursuing an MSc in Computer Science at University College Dublin (UCD).
Engineering Philosophy
How I think about building
From writing CLI utilities in 2021 to winning 10+ national hackathons, publishing AI research, and engineering production platforms at Norxio and ConnectWise. Every system started with the same constraint: a real problem with no time to waste.
I started programming in 2021, building terminal utilities—a CLI notepad, calculator, and contact directory—driven by a curiosity to understand how software works under the hood. When I started my Bachelor's in Artificial Intelligence & Machine Learning at SIES GST, I dove into competitive programming, ranking in the Top 20 of the CodeChef Turing Cup and curating contests for 500+ participants as Arena Coordinator.
My early role at HFC Enterprises reviewing Java and Python codebases established my discipline for clean architecture and production delivery. At EaseworkAI, I moved directly into applied agentic AI, orchestrating autonomous supplier-buyer procurement email workflows with CrewAI, Microsoft Graph API, and FastAPI to accelerate turnaround times by ~60%.
Hackathons became my primary testing ground: building FactFinder at Hackanova 3.0 TCET (which later was accepted and presented as a conference paper at Springer ICDSM 2024), winning 1st Place at The Great Bengaluru Hackathon with LoanSaathi, 1st Place at GajShield Hack 8 with Kavach.ai, and demonstrating SeaGuard to the Chairman of ISRO.
At Norxio, I served as Founding Software Engineer, designing B2B payment processing and internal dashboards with Next.js and Nest.js, deploying Cloudflare Workers microservices, and architecting an AI translation pipeline with Sanity CMS and Redis caching that cut external API hits by ~90% while maintaining a 100% SEO score.
At ConnectWise, I engineered an enterprise RAG dashboard across 10,000+ Jira tickets that cut manual triage time by ~75%, automated via GitLab CI/CD. At Cartman Labs, I shipped autonomous X/Twitter agents on Eliza OS and a Web3 community platform maintaining 99.5% uptime.
Today, alongside building the Eivora suite (eivora.app, apios.eivora.app, webaudit.eivora.app), I am pursuing an MSc in Computer Science at University College Dublin (UCD), focusing on Distributed Systems, Multi-Agent Systems, and Advanced Machine Learning.
01
How I Operate
01
Ship first, polish later
A working prototype in production tells you more in a day than a spec tells you in a month. Working software is the only true measure of progress.
02
Systems over features
A single script is a feature. An automated, fault-tolerant pipeline that ingests 10k records daily and surfaces actionable insights is a system. I build systems.
03
Constraints are the brief
24 hours, one GPU, strict latency budgets, zero room for failure. True constraints produce focused, robust software that survives contact with reality.
02