I work on the point where physical sensing meets AI — where the data comes from, whether it's trustworthy, and what breaks when a system leaves the lab. My current project is a wearable device for early Alzheimer's detection, with a UK registered design, now working through the electrode placement and signal integrity constraints that make it hard to build. I document the engineering as I go, including the parts that don't work.
About
I work in AI data quality in London — reviewing multilingual audio and conversational AI training data, including Indic languages. Doing that every day, I kept noticing the same thing: a model trained on averages fails the people who don't fit the average.
A regional accent isn't an edge case. A less-resourced language isn't an edge case. They're only edge cases if you built for the average. That observation is what the registered designs, the peer review work, and the software below all trace back to — checking whether a system still holds up when it meets someone it wasn't built to expect.
Quick facts
- Role: AI Data Quality Specialist, Sigma AI — London, permanent since April 2024
- Field: Data quality and trustworthy AI, applied to healthcare and biomedical systems
- Intellectual Property: 2 UK Registered Designs · 1 Indian Patent (published)
- Peer Review: Springer Nature (Scientific Reports) · SPACAID 2026 meta-reviewer
- Publications: 2 papers presented — PiCET 2026, IET Proceedings (Scopus-indexed)
- Education: MSc Global Project Management, University of Essex
- Education: BE Electronics & Communications Engineering
- Endorsing body: Tech Nation — Digital Technology route
Evidence
- Alzheimer's Disease Prediction Device with Artificial Intelligence (Registered Design · UK IPO) — Design No. 6521594 · Class 24, Medical & Laboratory Equipment · Registered 21 April 2026 · Granted 5 May 2026 · Co-authored with Shyamkumar Parikh. A registered design protects the device's appearance, not its function.
- Robotic Fleet Control Console (Registered Design · UK IPO) — Design No. 6521595 · Class 14, Data Processing Equipment · Registered 21 April 2026 · Granted 1 May 2026 · Co-authored with Shyamkumar Parikh
- AI-enabled smart healthcare wearable bracelet for continuous real-time patient monitoring and predictive health analytics (Published Patent Application · India) — Application No. 202611062464 · Publication No. IN202611062464 A1 · Filed 17 May 2026 · Published 17 July 2026 · Co-applicant Dr. Anurag Shrivastava. Published application — not yet examined or granted.
- Springer Nature — Scientific Reports (Peer Reviewer) — 3 manuscripts reviewed, 2026 · Reviewer certificate issued
- SPACAID 2026 (Meta-Reviewer) — IEEE-sponsored · IPS Academy · Invited via Microsoft CMT, 24 July 2026 · 5 papers assigned · 3 meta-review decisions completed — one Minor Revision, one Major Revision, one Desk Reject. Subject matter: deep CNN leukaemia classification, EEG-driven cognitive regulation, biomedical instrumentation, explainable multimodal deep learning for heart disease, FPGA collision avoidance.
- Adaptive Intrusion Detection System with Random Forest Classifier and Web-Based Interface (Presented · Co-author) — Ref. PU/PiCET26/COP/147 · Presented May 2026 · 8th Parul University International Conference on Engineering & Technology (PiCET 2026). With Mukesh Patidar, K. Mani Kumar, K. Shailandra Ganesh, T.S.V.N. Siva Prasad, Krupali Dave, M. Chaitanya Ram. Publication partner: IET Conference Proceedings, Scopus-indexed — proceedings pending.
- AgroSMART: An Advanced Agriculture Platform (Presented · Co-author) — Ref. PU/PiCET26/COP/153 · Presented May 2026 · 8th Parul University International Conference on Engineering & Technology (PiCET 2026). With Penta Mahesh, Gurram Naveen, Gujja Sai Lokesh, Krupali Dave, Poonam Songde, Rangala Vamshi Reddy. Publication partner: IET Conference Proceedings, Scopus-indexed — proceedings pending.
- Low Power IoT Architecture for Intelligent Electronics Applications (Published Paper) — IJTSRD · Nov/Dec 2022
- Design of Intelligent Embedded Electronics for Autonomous Smart Devices (Published Paper) — IJTSRD · October 2022
- IoT-Based Smart Monitoring System for Intelligent Electronics Infrastructure (Published Paper) — Web of Semantics · April 2024
- Smart Sensors Network for IoT-Based Electronics System Monitoring (Published Paper) — Web of Semantics · 2025
- AI Driven Fault Detection in Smart Electronics Power System (Published Paper) — IJARSCT · May 2026
Builds
- audio-qa-dashboard — Quality-tracking dashboard for multilingual audio and conversational AI data. Tracks approval rates, rejection patterns, and risk flags across projects, with a Random Forest model predicting which projects miss the quality gate early. (Data in the repo is synthetic, modelling patterns from daily QA review — no client records.). Live demo · Code
- alzheimer-ai-device — Companion repository to UK Registered Design 6521594 — signal processing and ML classification for early Alzheimer's detection, moving the concept from a design drawing into tested work. (Rebuilt on a publicly available clinical EEG dataset (OpenNeuro ds004504 — 88 subjects, 19-channel, CC0). ROC-AUC 0.789 under leave-one-subject-out cross-validation.). Code
- ai-care-alert — AI-assisted emergency alert system for people living alone, elderly, and disabled individuals — tiered alert logic and fall-detection simulation built around individual thresholds rather than one-size-fits-all defaults.. Code
- AI-IoT-Maintenance — Predictive maintenance using a Random Forest model on simulated IoT sensor data — data generation, training, and prediction with confidence scores.. Code
Contact