Soma Fellows Demo Day S24

The Soma Fellowship was born out of the desire to become more actively engaged with brilliant change-makers, doers, and thinkers as they work to fundamentally change the world we live in. We see it as a great opportunity to help some of the most brilliant founders at the outset with everything from storytelling to fundraising strategies and give them an edge to make change happen. 

We believe the next best thing to building the future is creating an ecosystem for the builders of the future to gather, learn, collaborate, and accelerate their ideas into the world. We’re excited to present our impressive cohort of brilliant teams in our S24 Fellowship below. If you’d like to get in touch with any of the teams, please fill out this form or click the button below.

Meet with the fellows

Meet the S24 Soma Fellows

We appreciate your interest and support and as always, if you know any founders / teams that would be a great fit for Soma don’t hesitate to reach out

Pulse automates supply chain procurement tasks with AI agents. Procurement teams spend 40% of their day emailing suppliers and updating spreadsheet data in ERPs, but with Pulse, they simply type in a purchase order ID and our AI agent follows up with all the suppliers and updates the ERP automatically (YC Launch)

Sid Manchkanti
Ritvik Pandey

We're building Playbook3d.com – a diffusion based render engine that reduces the time to final image with AI. It is accessible via web editor and API with support for scene segmentation and re-lighting, integration with production pipelines and frame-to-frame consistency for image and video.

Skylar Thomas
JD LeRoy

Enzo Health is a documentation co-pilot for post-acute care. Enzo Health's platform empowers post-acute care administrators by using artificial intelligence to facilitate their documentation review for maximum reimbursement and compliance.

Dan Conger
Zach Newman

Struction is automating the quoting process for construction suppliers and subcontractors - a $5 billion opportunity. We use proprietary computer vision models, multi-modal LLM's, and intuitive UI to bring the quoting process from days to within 15 minutes.

Joseph Rainey
Sam Tang

Maya is a regulatory workflow platform for REG teams that need to build better engagement between workflows.Teams can deploy and productionize a workflow, without the need for an RPA/process engineer.

Sat Ramphal
Sean Ramphal
Shivam Ramphal

Cocomo is building autonomous AI agents for marketing. We are targeting the $100B market spent yearly on marketing headcount and $50B spent on agencies.Our models automatically manage and optimize campaigns without human intervention by leveraging first-party data. Our team pioneered these systems at Starbucks, Amazon, Expedia, and Snap on $100M+ marketing budgets, seeing efficiency lifts of as much as 50%. We are advised by executives from Meta, Dropbox, Scale, Deloitte, and Bain to usher in a new way for companies to deploy marketing spend with AI.

Kunal Shah
Jack Chua

Airlines operate on razor thin margins, and already spend a third of their operating budgets on fuel. However as the industry starts to decarbonize and consume increasing levels of Sustainable Aviation Fuel (SAF), which is 2x-5x the price of jet fuel, airlines are facing the reality of balancing increasing fuel costs against growth objectives. AeroFuse is offering the industry an alternative. By harnessing the power of fusion - which is millions of times more energy dense than fossil fuels - AeroFuse is developing fusion propulsion systems to help airlines reduce fuel burn and emissions by 30-100%, whilst lowering fuel costs, and increasing aircraft range by 20-30%.

Our vision is to bring zero-emissions propulsion to large aircraft - without the green premium. AeroFuse's core innovation is in the development of a fusion heat exchanger technology to efficiently extract heat energy from fusion plasmas and integrate this process into existing gas turbine engines. Our technology will be available as a retro-fit and line-fit solution, and our approach maximizes the re-use of existing turbine technology and existing tube and wing aircraft architectures.

Rama Myers
Ryan Weed

At Fasten, we are introducing the first-ever consumer credit card that rewards points for auto loans. The card offers accelerated points for daily transportation-related purchases such as gas, insurance, tolls, and parking and one point for everyday purchases. Points can be redeemed within the specific dealership group the consumer is connected to and for travel and gift cards.

Additionally, we will offer a merchant processing product for dealers, enabling an "on-us" payment model for Fasten transactions at the dealership. This setup allows us to offer dealers a ~25% reduction in payment processing fees when customers use the Fasten card compared to other credit cards.

Our goal is to use the app to promote dealer and third-party products, i.e., come test drive the new Hyundai Santa Fe for 500 Fasten Points, or if you purchase your auto insurance here, the cardholder will receive 1,000 Fasten points.

Fasten uniquely integrates its credit card rewards system with dealership reinsurance programs, enabling dealers to use accrued points as investable assets within their insurance captives. This innovative approach allows dealerships to enhance financial flexibility by transferring revenue from a high-tax entity to a low-tax entity.

Jacob Zachs

Mind Company is an EEG foundation model business developing high-accuracy linguistic neural interfaces by decoding text from modern, noninvasive, EEG sensors. Founded by Rayhan Ahmed, Mind Company is a dedicated group of researchers and engineers from Mach Industries and Bridgewater Associates with the mission of bringing products to market that can reliably and safely interpret human thought. Here is our memo for a deeper look.

Rayhan Ahmed

Tilde is revolutionizing how we understand and control foundation models by making them transparent and provably reliable. Models currently are black boxes, which is not only a barrier to ensuring safety but also unlocking performance. Our interpretability agent 1) dissects your model, 2) identifies when and why your model fails, and 3) gives you the power to precisely edit its behavior to eliminate those failure modes. Instead of the traditionally unreliable, time-consuming, and costly methods, both human-in-the-loop and automated model calibration become effortless with Tilde. Through proper interpretability, Tilde is creating the future of 0% error case foundation models.

Dhruv Pai
Benjamin K
Mason Wang