Projects
Selected work across agents and automation, search and knowledge systems, multimodal and document intelligence, and evaluation and responsible AI. The examples connect hands-on research with reusable capabilities and cross-team delivery.
Project Portfolio
Search & Knowledge Systems
Oracle Cloud Infrastructure
Tech Lead
2025
- Built and grew applied-science teams delivering enterprise semantic layers, personalized search and agents, and support automation.
- Architected shared knowledge infrastructure combining multimodal and multilingual retrieval, ranking, knowledge graphs, and access-aware grounding.
- Co-first author of hard-negative mining research for domain-specific enterprise retrieval.
- The reported cloud-services test-set comparison achieved MRR@10 of 0.64 versus 0.57 for ADORE+STAR; this is a study result, not a production business metric.
Evaluation & Responsible AI
Oracle Cloud Infrastructure
Research and evaluation
2024-2026
- Co-author of AccessEval, a study of disability bias in large language models recognized with the EMNLP 2025 Social Impact Award.
- First-author research on multimodal evaluation examines visual reasoning and sensitivity to changes that preserve meaning.
Agents & Automation
Oracle Cloud Infrastructure
Applied science and technical leadership
2023-present
- Work on support agents, ticket automation, and incident triage and resolution workflows for Oracle Support, connecting research questions with product requirements.
- Architect Agent Hub routing, skill/tool retrieval, and shared context and memory for long-horizon tasks.
- Developed a personalized leadership-coaching agent pilot combining assessment-driven plans, persistent memory, retrieval-grounded recommendations, role-play practice, and proactive nudges.
Multimodal & Document Intelligence
Oracle Cloud Infrastructure, Jio Platforms
Applied science and technical leadership
2020-present
- Owned applied-science contributions to multimodal and agentic enablement of Oracle Fusion ERP and Oracle AI Database 26ai, spanning multimodal RAG, embedding fusion, image understanding, and text-to-image capabilities with product engineering.
- Developed document-understanding capabilities spanning OCR, key-value and table extraction. Earlier at Jio Platforms, led computer-vision work including video analytics, detection and tracking, and OCR-based document understanding.
- Research contributions include graph models, low-data learning, and synthetic document generation.
For collaboration or technical leadership opportunities, get in touch.