01
Confidential
Aug, 2026 - Present
Software Engineer - Full Stack (AI)Remote, Miami
- •Architected and implemented an A storage design where every object records who owns it, so access is checked per object instead of trusted by default. across the platform’s object storage, closing a cross-tenant security gap, and shipped a per-workspace A workspace-level place to keep files that are not tied to any single task. on top of it
- •Led a major architecture revamp of The step where a person edits generated results before they are finalized. end to end, changing how it works under the hood and making sure Every prior edit stays recoverable, even after the workflow that produced it runs again. holds up across a workflow re-run, reviewed and merged
- •Introduced self-serve Lets each customer show their own name, logo, and colors instead of the platform’s default. across the platform, letting owners update branding information, and centralized report state and rendering into a single system used across the app
02
- AI-Powered eKYC & Customer Onboarding Platform
Sep, 2025 - Jul, 2026 (10 mos)
Software Development EngineerOnsite, Bangalore
- •Part of the HyperVerge Onehyperverge.co/in/hyperverge-one/ platform and Secure identity verification and KYC workflow platform for generating verification links, onboarding users, validating documents and faces, and retrieving compliance-ready API results. product teams, building core infrastructure and developer tooling for a platform HyperVerge publishes as having processed 2B+ identity verifications for 500+ enterprises across India and the US. The platform serves close to 200 million requests a day
- •Designed and built a A config-driven mechanism serving per-client interface values at runtime. spanning the platform's application and API layers, a single config-driven mechanism that lets enterprise clients customize interface text without a code deployment
- •Redesigned the platform's The client-facing reporting view for usage and verification data. end to end, owning both the product direction and the implementation, and extended it with the same per-client customization the label configuration system introduced elsewhere on the platform
- •Led Testing a service under heavy simulated traffic to see how it holds up before real users do. across major services, alongside UI/UX and feature improvements shaped by product and client direction
- •Built the end-to-end testing pipeline to stay dependable as more of the codebase shipped through an More of the codebase written by AI coding agents rather than by hand, which raises the bar for what a test suite has to catch., lifting engineering velocity by roughly 50-60%
- •Built three internal Automation triggered from and reporting back into Slack, the team's day-to-day tool. used across the company's roughly 400 people: trial-expiry alerts that kept renewal conversations and payment on schedule, a A defined request-and-sign-off path, routed straight to the right owner. for platform access that cut turnaround from days to minutes, and a request tracker giving the team one accurate, trackable record of every bug, feature request, and client ask across Jira and Google Sheets
- •Led a cost-optimization pass on the platform's logging infrastructure, cutting spend by roughly $1,800 a month
- •Resolved client-impacting production outages on tight timelines by tracing issues across multiple microservices to their root cause
- •Built production-ready web applications on pyOpenMSOpen-source Python bindings for the OpenMS mass-spectrometry library, used for proteomics and metabolomics data analysis workflows.pyopenms.readthedocs.io/en/latest/, used by thousands of researchers across the proteomics community
- •Shipped the Peptide m/z ToolOpenMS web application for peptide mass-to-charge calculations with mzML file processing and data visualization.github.com/OpenMS/mz-calculator, a full-stack tool for mass spectrometry calculations with mzML file processing and data visualization
- •Set up automated Continuous integration and deployment: every change automatically tested and shipped through to production. with GitHub Actions and containerized deployment via Docker
- •Worked directly alongside the The team steering OpenMS, the open-source mass-spectrometry library used across the proteomics research community. at ETH Zurich, the University of Tübingen, and the University of Toronto
CI/CD
04
Stealth Startup
Dec, 2024 - Jan, 2025 (1 mo)
Frontend DeveloperRemote, India
- •The only frontend engineer on the team; every screen in the product went through this role., working from shared Figmawww.figma.com designs, from the landing pages through the full dashboard
- •Partnered directly with the founding team on design and implementation decisions as the product took shape
05
May, 2024 - Jul, 2024 (2 mos)
Machine Learning Trainee, Unnati Industrial Training ProgramRemote
- •Built a machine learning model for Spotting when another vehicle merges into a car's lane closely enough to be a collision risk.
- •Orchestrated a YOLOv8A real-time object detection model, used here to recognize vehicles in the frame.docs.ultralytics.com/models/yolov8/-based detection system, achieving 80% precision across diverse road conditions
YOLOv8Machine Learning
06
Symphony Music
Mar, 2020 - Sep, 2022 (2 yrs 6 mos)
Lead Developer TeamRemote
- •Architected and scaled a high-performance Discord music bot serving 300,000+ users across 300+ servers
- •Implemented 40+ real-time audio filters and multi-platform streaming from YouTube, Spotify, and SoundCloud
- •Maintained and shipped continuous performance improvements over 2+ years with Every release went live without interrupting the bot for users already connected.
Love what I do? Let's talk!