Denzil Nelson
denzil10.github.iodenzilnelson114@gmail.comLinkedIngithub.com/Denzil10x.com/denzilnelsonai
AI engineer. I build AI agents that run unattended, and the checks that catch them when they lie: restricted permissions, claims that must name the live check behind them, and verifiers the agents can't edit. Open to AI engineer and forward deployed engineer roles, hybrid in Navi Mumbai, or relocating within India or abroad.
Selected work
3 lanes of agents research, build, list and promote small products on a schedule with nobody in the loop: 582 scheduled runs over 56 days, 13 products live. Deterministic gates before every model call, a verifier after it that the agents can't edit, and every public number generated from the fleet's own logs.
Designed and ran a GitHub onboarding program where newcomers learn open source by shipping real pull requests: 180 stars, 109 forks.
My agents got more reliable when I gave them less to do · 88% of my coding agent's usage goes on reading · Letting agents drive a real browser on the Mac I'm typing on · A prompt is not a control: two autopilots that forged their own evidence
Experience
Employer details on request.
Built a responsive, minimal site for a gaming studio, refined its UI from team feedback, and set up hosting and deployment for the client.
Built a banking website prototype with real-time transaction updates (AJAX, PHP). Received a letter of recommendation.
Publications
IEEE Xplore; presented at INCOFT23
Journal of Electrical Systems (Scopus-indexed)
Earlier projects
Mobile app for fishermen: ResNet fish-species identification, sea forecasts from the Windy API, and YOLOv8 catch tracking on vessels.
CNN for cervical fracture detection in CT scans, 86% accuracy; showed coronal views work for the task with transfer learning (EfficientNetV2B3, ResNet).
Online-abuse analysis with Hugging Face NLP models over scraped posts, with a lookup of the nearest cybercrime cell.
Education
B.E. Computer Engineering, Honours in Data Science, CGPA 8.5/10
Leadership and recognition
Built AR lenses with over 1M views.
Ran technical workshops and events.
Also completed Hacktoberfest 22.
Skills
AI agents · LLM evaluation · Python · Browser automation (Playwright) · Prompt and context engineering · Deep learning (TensorFlow, YOLO) · Computer vision · NLP · JavaScript, React, Node.js · MongoDB, MySQL · AWS · Git and GitHub