SELECTED WORK / 01
A selection of enterprise AI, agent products, and physical AI work. Some production systems are internal, so the descriptions focus on their purpose and architecture.
Private models, document ingestion, indexing, retrieval, and agents connect organizational knowledge with internal systems. People can search documents, query structured data, and produce grounded reports.
Private LLM · RAG · vLLM · AgentsAn AI-native content operating system for creators and small teams. It plans from a brief and prior performance, prices the work, creates images, video, voice and music after approval, then publishes and learns from results.
Explore SoeMind ↗A reinforcement learning driving policy trained in CARLA with semantic segmentation, then transferred to a computer-controlled electric golf cart with steering, throttle, and braking.
PPO · CARLA · Jetson · Semantic segmentationA depth-camera system for water-pipe trench quality control. It processes point clouds into trench profiles and measurements and runs on an NVIDIA Jetson Xavier NX for field use.
View research project ↗These projects form a connected internal AI environment rather than isolated demos.
Turns fragmented notes, documents, and spreadsheets into structured investor intelligence: interactions, milestones, status, and next actions.
Tracks company activity histories and risk indicators, with dashboards and AI-generated summaries.
Connects employee profiles, skills, interests, roles, learning opportunities, and skill gaps to support career development.
Employee profiles, leave and calendar workflows, and process automation, built as a foundation for future skills intelligence.
My earlier work covers computer vision, databases, generative AI, NLP, and automation. The original project notes remain available in the archive.
Browse earlier projects such as road-scene segmentation, augmented reality, text-to-image applications, customer-intent detection, yoga-pose classification, and autonomous transport.
Open earlier project notes ↗