Tools born from my own passions, built by breaking big ideas into simple steps.
Paste a YouTube link, get six instrument stems you can mute, slow down, transpose and loop. CloudFront serves the page, which sends the link to a Lambda API that queues a job in SQS. An on-demand EC2 worker splits the audio with Demucs into S3, and the browser caches the stems and plays them in sync with Tone.js.
A CNN that turns patches around keypoints into feature descriptors, trained without a large dataset. Crops and augmentations of a single source image are paired so the network learns which patches show the same point. Keypoints from OpenCV are described, matched across image pairs and benchmarked against SIFT: within 1.5% of its accuracy over 70 pairs, with 4% more matches.
At ML Research Labs, Trellis Data's R&D division, I was part of their initiative to preserve indigenous languages through machine translation and low resource learning. I worked across their early-stage proof-of-concept LangGraph agentic translation workflow, inspired by the Machine Translation from One Book paper. A new chunking strategy I introduced improved translation accuracy by 20%, and rewriting LLM queries to target specific recurring errors, such as missed pronouns, improved translation quality by 8%. Switching to persistent FAISS vector stores also made load times over 80% faster. I then built a FastAPI interface so external users could interact with the workflow, and parallelised the evaluation setup with asyncio to run 20+ Gemini mixed-model setups, cutting evaluation time by 36% while keeping translation quality within 3%.
I worked in HPE's HPC division on the System Monitoring Application (SMA), a distributed monitoring and diagnostics application for HPC systems, used by several of the world's top supercomputers. My responsibilities included extending SMA's Kafka-based data collection services, building visualisation dashboards in Grafana, and tracking down issues in our production test environments ahead of each release. My proudest accomplishment was being picked as the sole representative from a global team of senior engineers to fly out to LUMI, Europe's largest supercomputer, and fix an issue that was blocking acceptance. Over one week on-site with the customer's engineers, I found the right data filtering setup to get critical monitoring data flowing in without gaps under a tight ingestion limit, and saw SMA through to sign-off. While I was there, I also redesigned 50+ Grafana dashboards with summaries and insights shaped by discussions with the on-site hardware engineers who keep the system running, and made them up to 80% faster to load by optimising the underlying SQL queries.
I started at HPE as an intern in its largest R&D division in India, working on their data centre digital transformation initiative. The goal was to replace fragmented Excel-based tracking, for which I built a full-stack MERN inventory management application with a three-tier RBAC approval workflow. It also had an in-app migration feature for legacy data from FY 2019-20 and FY 2020-21, with MongoDB schema and Formik validation flagging malformed values for manual updates. The app enabled real-time monitoring of 5,000+ consignments a year, and turned my internship into a full-time Software Engineer offer, with a recommendation from my mentor.
My first internship was in Fourth Signal's web development team, enhancing their in-house, client-facing platform. Through discussions with my mentor, I refined the methodology for a quantitative mutual fund risk-score model, ranking 50+ debt and hybrid funds. I implemented the finalised model and its front-end components, delivering it as a fund-ranking and recommendation feature on the platform.
I enjoy learning how things work, and I'll happily go down a rabbit hole to find the most efficient way of doing something. I'm always keen to give something new a go, though what I enjoy most is picking a few things and mastering them.
Outside of work, I'm aiming for a 500kg powerlifting total and, one day, to play a full song by my idol, Alexi Laiho. I like having a concert to look forward to, and I've met or seen all five of my top artists live. On quieter days, I enjoy riding my bike, car spotting and listening to podcasts. If you're looking for one, I'd highly recommend World War II with Tom Hanks.
Do reach out!