Participated · Series B · Dec 2023

Totus Medicines discovers and develops small molecule medicines using a novel DNA-encoded covalent library technology combined with artificial intelligence/machine learning. Its platform can screen billions of drug candidates against thousands of targets simultaneously, enabling discovery of molecules that may be superior to those found with prior technologies, including candidates for previously undruggable targets. The company plans to advance its clinical program, expand its pipeline, and evolve its platform. Totus closed a $66M Series B financing to support those efforts. The round was led by DCVC Bio with participation from North Pond Ventures, Camford Capital, and the Regents of the University of Minnesota. The company also announced Nassim Usman as President & CEO while co-founder Neil Dhawan will transition to CSO and Head of R&D to continue overseeing the platform, programs, and data. Totus Medicines, a Massachusetts-based biotech, exited stealth after raising $40 million in a Series A to advance PI3Kα-targeted therapies. Its lead candidate, TOS-358, produced tumor regressions in preclinical models of breast, colon, lung, and stomach cancers where other PI3Kα inhibitors showed only mild effects. The company plans to begin clinical testing of TOS-358 in the second half of next year. Totus uses a drug-discovery platform that employs molecular tags to track drug binding in individual cells, enabling screening of billions of molecules across thousands of genes in parallel. The company positions its asset as an improvement on existing PI3Kα inhibitors, citing limited success in under 10% of patients with PI3Kα-mutant cancers. Totus is led by CEO and co-founder Neil Dhawan, Ph.D., who previously co-founded Dual Therapeutics.

Participated · Series B · Jul 2023

SURGE Therapeutics develops intraoperative immunotherapy using its SURGERx™ platform and an injectable biodegradable hydrogel to concentrate immunotherapy at surgical tumor resection sites. Its lead candidate, STM-416, is in a Phase 1/2a trial for recurrent bladder cancer, and the company recently dosed the first two patients. The intraoperative approach is intended to provide extended, localized release of immunotherapy to prevent tumor relapse and induce systemic antitumor immunity. In multiple aggressive murine models, intraoperative immunotherapy markedly improved survival relative to traditional systemic or local administration. SURGE plans to advance multiple clinical programs, expand the team, and generate proof-of-concept event-free survival data with its lead program. The company is based in Cambridge, MA. SURGE Therapeutics is developing an injectable biodegradable hydrogel (the SURGERx™ platform) that enables extended, localized release of cancer immunotherapy at the site of surgical tumor resection. The technology was built on research from founder and CEO Dr. Michael Goldberg's laboratory at Harvard Medical School. In multiple aggressive murine models, intraoperative delivery via the hydrogel vastly improved survival benefit relative to traditional systemic or local routes. SURGE seeks to reprogram the post-surgical immune response from immunosuppressive to immunostimulatory to destroy residual local and distal cancer cells and reduce recurrence. The company plans to use recent funding to accelerate development of its intraoperative immunotherapies, expand the team, and initiate clinical trials. SURGE is based in Cambridge, MA and has assembled a team of operators and advisors across localized therapeutics, translational immunotherapy, surgical oncology, and drug delivery.

Led · Seed · Jul 2023

FedML offers an MLOps platform that lets companies and developers train, deploy, monitor and improve AI models on cloud or edge devices while enabling collaborators to share data, models and compute resources. The platform can run custom models and open-source models, and users can create collaborator groups that sync AI applications across devices and add servers or mobile devices to contribute compute. FedML recently launched FedLLM, a training pipeline for building domain-specific LLMs compatible with libraries such as Hugging Face and Microsoft DeepSpeed. The company also aims to build a community of CPU and GPU resources to host and serve models, proposing incentives such as tokens or other compensation for contributors. Financially, FedML raised an $11.5M seed round at a $56.5M valuation and reports a total of $13.5M in its war chest inclusive of the new funding. The company has a 17-person team, about 10 paying customers including a tier-one automotive supplier, more than 3,000 users globally, and reports over 8,500 training jobs across more than 10,000 devices.

Led · Seed · May 2023

Lavita AI operates a decentralized, patient-driven health information marketplace hosted on the Theta Metachain that combines privacy-preserving computation with advanced AI. The platform returns ownership and control of health data to individuals and enables them to securely share and monetize their data via LAVITA TNT-20 tokens. Lavita’s product roadmap includes clinical-trial matching, convening data cohorts for large-scale genome- and phenome-wide studies, secure multi-party data sharing for drug discovery, and tools for health institutions to train and fine-tune domain-specific LLMs. The company is integrating Theta Edge Network for distributed computation and storage to support healthcare model training and a custom Theta subchain for private clinical-trial matching. Lavita closed a $5M seed financing to accelerate development of its AI-first platform and applications for patients, providers, and research institutions. The company was selected for the Microsoft for Startups Founders Hub Program and has published a white paper describing its approach.

Participated · Seed · Dec 2022

Tahoe Therapeutics provides AI-powered models of the human cell that generate large-scale, perturbative single-cell datasets to inform drug design. The company’s platform aims to enable a new generation of biological foundation models for drug discovery. Tahoe plans to generate one billion single-cell datapoints, mapping how tens of thousands of drug molecules interact with human biology. It intends to use the new funding to advance its therapeutic programs toward the clinic and to launch a new model of strategic collaboration. Led by CEO Nima Alidoust, Tahoe was founded by a team of scientists and technologists advancing drug discovery, genomics, and machine learning. The company is based in South San Francisco. Vevo Therapeutics develops an in vivo drug-discovery platform, Mosaic, that combines scalable in vivo data generation with AI models to identify drugs for more patients. The platform captures in vivo disease context at single-cell precision and aims to better represent patient diversity versus current in vitro assays. Mosaic builds on technology developed by co-founders Hani Goodarzi and Johnny Yu and is exclusively licensed from UCSF Innovation Ventures. The company is led by CEO and co-founder Nima Alidoust, Ph.D., alongside co-founder Kevan Shokat, Ph.D. Vevo plans to use recent financing to accelerate growth and expand operations. The company is based in San Francisco, CA.