Timeline
The path so far.
From a national-level taekwondo black belt in Goa to shipping production voice AI and writing twenty-two research papers on the same systems: one published at ICANN 2026, nine more submitted. Newest first.
2026
Brillianse — 13-app academic ops platform, live
VoiceQubeMerged 447 of the repo's 2,688 pull requests and wrote more commits than anyone else on the team. Thirteen apps in one monorepo (8 Next.js, 5 Expo), a Hono API with 152 route modules, Postgres with 323 tables. Live for a 12-campus college group: 4,735 students, 137,821 exam submissions, 7.6M answers recorded. Built the admissions CRM with an AI voice agent, and AI grading of handwritten answer sheets. Launched publicly as Brillianse AI by Telangana's IT & Industries Minister in October 2026 (The New Indian Express). Case study.HotelClaw.ai — integration layer of a multi-tenant hotel AI platform
Side projectOracle OPERA exports and an Opera Cloud (OHIP) adapter generalised into one PMS sync runner, with Mews, Cloudbeds and Apaleo adapters on top; LINE and WhatsApp ingestion feeding an LLM intent router that leaves every reply to staff; tenant isolation made a CI check. Case study.Left VoiceQube after three years forward deployed
VoiceQubeThree years embedded with US and Indian clients: six client products (Altfolio, Kissan Pride, Finie, Northway Tech, and the sister products Phiny.ai and Crosslane) and three in-house ones (CueKit.ai, Unifund and Brillianse AI).Adaption — AutoScientist winner, now an ambassador
CommunityWon 1st place ($4,000) in Data & Visualization at Adaption Labs' AutoScientist Challenge, and since September 2026 a volunteer ambassador for the non-profit: running online and offline events, testing new formats and platform features, and turning member feedback into changes. The entry.Co-founded CELABE
Side projectCo-founded an AI-powered wedding & events platform. Architected end-to-end across iOS, Android, and web; live on the App Store and Play Store. ML-powered planning, vendor marketplace, viral QR photo-sharing.Twenty-two research papers · one published at ICANN 2026
ResearchThe systems-level work behind the products. Latency Optimization in Production Voice AI Pipelines (−41.8%) is published at ICANN 2026 (peer-reviewed, Springer LNCS, doi:10.1007/978-3-032-38404-1_36). Also under review: Personality & LLM Trust, the SSIM line-art audit, Done Is Not Correct (agentic CAD reliability), TRIBE v2 Predicted-Cortex Features Beat an Unmatched V-JEPA2 Video Baseline for Memorability on VideoMem and Lose on a 499-Clip Memento10k Subset, and two submitted in October 2026 (PacificVis, International Journal of Data Science and Analytics). Full list with DOIs and ORCID on the research page.Certificate of Reviewing
ServiceReviewed a manuscript for the European Journal of Computer Sciences and Informatics — contributing to research quality on the other side of the page.Merged PR into everything-claude-code
Open sourceContributed to one of the fastest-growing AI repos (236k+ stars, 35k+ forks). Started life as zenith.chat — built with Claude Code in ~8 hours, won 1st place at the Anthropic × Forum Ventures hackathon (NYC, 2025).Shipped SmokeMukti & Oravira to the App Store
Health tech
2025
Started research
ResearchBegan formalizing the production work into papers — voice AI latency, intent detection, MCP architecture, LLM optimization, trust in LLMs, and multi-agent drift. The voice AI latency study (−41.8%) was accepted at ICANN 2026 and is now published in the Springer LNCS proceedings.Adopted Claude Code in production
AI devWired Claude Code into the team's daily build / test / ship loop; event-driven agent orchestration across the team's repos.Real-time AI avatars
Built bi-directional voice + video + data avatar experiences with LiveKit + WebRTC.AI for Marketing — BITS Law, Mumbai
Speaking3-hour technical workshop: creative AI tooling, content + distribution automation, zero-code CRM pipelines, hands-on with LLM and image-gen tools.
2024
VoiceQube — leading teams, shipping products end-to-end
Team leadThree years leading engineering teams: 30–40 engineers on and off across four product lines, through architecture reviews, code reviews and mentorship, forward deployed with overseas clients.Client AI products, end to end
FDEShipped Phiny.ai (AI career-prep with real-time voice mock interviews — Pipecat + LiveKit + Claude + ElevenLabs + Deepgram, 516 TS files / ~101k LoC NestJS, ~17k LoC FastAPI), Crosslane (career counselling, 476 TS files / ~107k LoC across 13+ modules), and Northway Tech (AI voice caddie featuring 2× US Open champion Andy North).Real-time backend at 40,000+ locations
NestJS + AWS architecture with event-driven workflows, MongoDB geospatial indexing, Redis caching, AWS Lambda + MediaConvert media pipelines, CloudFront edge delivery.Co-founded V3NEX, a non-profit
Side projectIntroduced kids to Generative AI — workshops on text and image generation with local AI models, mentored students through deployed projects. v3nex.vercel.app
2023
Joined VoiceQube — Senior Software Engineer, Forward Deployed & Team Lead
The start of a three-year run shipping production voice AI for US clients, and of leading the teams that built it.Best Outgoing Student
AwardDon Bosco College of Engineering, Goa. CGPA 9.28.Gold at University Games — Taekwondo
State level (again).
2022
Chairman, C-Codes — INSPIRUS 2K21
Organised a two-day national tech event with 300+ participants and three sponsors (Kilowott, OneShield Software, FIIRE).Royal Finance — Slovakia
React frontend revamp + admin panel for a Slovakian fintech startup (royaltyfinance.cz).Smart India Hackathon — Grand Finale
Built Coco.Techgium by L&T — Grand Finale
Built ParkMe.Pai Kane Group — ServeME
A React Native app and admin website for the group; full one-life-cycle internship project.Kusol Labs — internship
Built a web app with React, Material UI, MongoDB. Learnt scrum and agile.Sessions on React & React Native
Walked first- and second-year students through React, then a follow-up on building chatbots with Dialogflow + React Native.
2021
Started exploring AI / ML
Joy of Computing Using Python (IIT Madras), Machine Learning for Trading specialization (Google Cloud × NYIF).Web dev at scale
Shipped multiple full-stack projects across React, Node, MongoDB. Found my voice in TypeScript.