Designing and deploying production LLM, multimodal, and agentic AI systems — from data pipelines and model evaluation to sub-3000ms inference at scale in healthcare and enterprise.
I am currently an AI Engineer on the Jio ADLC Team, where I engineer an autonomous multi-agent software-delivery platform with persistent workflow state, MCP services, security gates, and human-in-the-loop governance.
I'm a Data Science and AI Manager based in Navi Mumbai, India. My focus is on production AI: large language models, multimodal systems, agentic workflows, and the MLOps that keeps them reliable in the real world.
Previously, I worked as an AI Engineer at Jio Partner World (Jio Platforms Limited), where my team developed agentic AI pipelines and real-time multimodal verification systems serving 100,000+ daily Fiber Device installations.
Before that, I led healthcare AI at Jio Health Technology — shipping LLM-based clinical summarisation and a personalised recommendation system for 2M+ active users. My early career was as a research analyst in mental health at the University of Ghana, where I built statistical models and dashboards for national health policy.
I hold an MBA with a specialization in Analytics from the University of Massachusetts Global, USA, a Post-Graduate Programme in Data Science and AI from Jio Institute (First Class) in India.
Designed a three-agent architecture comprising Orchestrator, Delivery, and Verification agents with separation of duties, least-privilege tool access, and explicit permission boundaries.
Implemented stateful orchestration, eligibility checks, and work locking to support idempotent, replay-safe execution and prevent duplicate processing of Azure DevOps events and work items.
Integrated MCP-based services for persistent context and state, build execution, notifications, visual verification, and security scanning.
AI Engineer
Jan 2025 – Sep 2026
Jio Partner World, Jio Platforms Limited
Designed and deployed agentic AI systems automating end-to-end ML workflows — data ingestion, annotation, evaluation, benchmarking, fine-tuning, and real-time inference — reducing manual intervention and improving reproducibility.
Built a multimodal AI system combining computer vision and LLM integration for device verification, improving onboarding accuracy from 89% to 99% and eliminating repeat field visits across 100,000+ daily AirFiber installations.
Developed real-time inference architecture with 3,000ms latency on Azure and GCP, supporting high-availability production operations at scale.
Manager, Data Science and Business Analytics
Nov 2023 – Jan 2025
Jio Health Technology, Jio Platforms Limited
Built an LLM-based NLP system for automated medical report summarisation, reducing hospital discharge processing time by 60% and improving clinical workflow efficiency.
Developed a personalised recommendation system leveraging behavioural and clinical data for 2M+ active users, increasing engagement by 35%.
Designed evaluation and validation pipelines for healthcare AI aligned with clinical standards, enabling confident adoption by medical professionals.
Research Analyst (Data Science)
Jun 2019 – Jul 2022
University of Ghana, School of Nursing — Department of Mental Health
Conducted statistical analysis and ML modelling on national healthcare datasets to support evidence-based policy and public health planning for government stakeholders.
Built interactive dashboards in Tableau and Excel to visualise healthcare service delivery trends and patient outcomes.
Integrated multi-source datasets via APIs and SQL pipelines to create unified analytics supporting public health interventions and resource allocation.