As the Chief Technology Officer at Datum Technologies Corporation, an Avenue Capital subsidiary, I scaled and managed a team of five engineers to deliver custom commercial real estate (CRE) software solutions for mid ($50 MM+) to large-sized CRE ventures.My technical leadership led the reverse engineering of Argus, the leading commercial real estate financial modeling software. I enhanced it with AI features to update and modify cash flow projections, working closely with clients to meet their needs.I led the development of Retrieval Augmented Generation (RAG) pipelines using Langchain, Google Vertex AI, and Gemini 1.5 Pro to extract financial data from leases and update cash flow projections for real estate portfolios.I also built a Langchain LCEL chain with Pinecone vector store for accurate document-based Q&A retrieval, integrating with ChatGPT and Gemini LLMs using flexible embedding swaps.To support these innovations, I developed a serverless worker node using Runpod.io for memory-intensive computations, containerized with Docker and maintained through a CI/CD pipeline. The solutions were built with a modern tech stack including TypeScript, NextJS, TailwindCSS, Firebase, and Langchain, ensuring robust and scalable outcomes for our clients.

Makeayo is the easiest way to create generative art straight from your PC.At Makeayo, I built a consumer desktop application for running Stable Diffusion using local hardware. Out-of-the-box features include text-to-image and image-to-image transformations, inpainting, multi-ControlNet, tiling, support for LoRAs and LyCORIS, HD upscaling, auto-updating, and model browsing.Currently, Makeayo generates over 1,000,000 images weekly.Makeayo is built with Electron, Pytorch, Fast API, React, and TailwindCSS.

As the Founder of Tenami, I spearheaded the development and launch of an innovative premium content marketplace tailored for livestreamers and influencers. I successfully raised capital to bring this vision to life, securing sublicensing agreements with top-tier studio television providers, granting access to over 10,000 studio films and TV series for distribution on platforms like Twitch and YouTube.I led the technical architecture, building a scalable video-on-demand streaming server that leveraged HLS and MPEG-DASH for adaptive bitrate streaming. This system ensured smooth delivery of content, including Paramount productions, encoded in AV1 and H264 via FFMPEG, stored securely on AWS S3, and delivered through Bunny CDN.I also built a Chrome extension that synced Netflix, Max, Disney+, and Crunchyroll with Twitch, handling 20,000+ concurrent users, and developed the Tenami desktop app with Widevine DRM for secure content delivery. My efforts positioned Tenami as the host of the world's largest Netflix watch parties.

4th hire for Niche Biomedical, in charge of leading the hardware design of ExaStim, a wearable multi-channel transcutaneous spinal cord stimulation device designed to help rehabilitate paralysis and loss of function due to spinal cord injury.I developed the schematic and layout for the entire wearable device, including power tree, trans-impedance amplification, and communication. The device can deliver up to 20 watts (+-100V, 200mA) through the human body using preset stimulation waves and patterns controlled through a Bluetooth companion app.Responsibilities also included board bring-up, validation, and regulating system design specifications to FDA standards.The device is currently used in the ASPIRE study for early feasibility trials across the United States, including at Mt. Sinai Hospital and Spaulding Rehabilitation Hospital (Harvard).

Recipient of the Annenberg fellowship, USC's largest fellowship offering for Viterbi students including a 6-year funding package. I attended for one academic year before leaving to join Niche Biomedical.

Interned at the Vision Research Lab at UC Santa Barbara as part of the BisQue platform:"Store, visualize, organize and analyze images in the cloud. Bisque was developed for the exchange and exploration of biological images. The Bisque system supports several areas useful for imaging researchers from image capture to image analsysis and querying. The bisque system is centered around a database of images and metadata. Search and comparison of datasets by image data and content is supported. Novel semantic analyses are integrated into the system allowing high level semantic queries and comparison of image content."