Problem and Solution Report: Synthetic Sciences

Synthetic Sciences addresses a critical bottleneck in modern scientific research: the lack of end-to-end automation and the difficulty of generating high-quality process data for training autonomous systems. Traditional computational research workflows are often fragmented, requiring scientists to manually bridge the gap between literature review, model training, experiment design, and documentation. This manual process is not only slow but also leads to issues with reproducibility and a lack of structured data that could be used to improve research efficiency through machine learning.

The company's core thesis is that truly capable AI scientists require two components that are currently missing: a human-centric product suite that generates rich process data and the research infrastructure necessary to transform that data into autonomous systems. Without a platform that captures the 'how' of scientific discovery—the intermediate steps, the failed experiments, and the reasoning—it is impossible to train Reinforcement Learning (RL) environments that can eventually perform research autonomously.

Synthetic Sciences' solution is a platform of 'AI co-scientists' designed to run research end-to-end. The product allows researchers to delegate complex tasks such as literature reviews, model training on cloud GPUs, and the generation of publication-ready LaTeX documents. By providing specialized 'modes' for biology, research, and model fine-tuning (Flywheel), the platform integrates into the scientist's existing workflow while simultaneously acting as an orchestration layer for GPU resources and credentials (e.g., GitHub, Hugging Face).

The value proposition lies in the creation of a continuous workspace where agents built for scientific execution can operate persistently. This approach not only increases the immediate productivity of individual researchers but also builds the foundation for agentic scientific workflows. By combining a user-friendly product with RL environments and process-based training data, Synthetic Sciences aims to move the field toward a future of fully autonomous, reproducible, and scalable scientific discovery.

Market and Competitors Report: Synthetic Sciences

Synthetic Sciences operates in a rapidly evolving landscape where artificial intelligence is being integrated into every stage of the scientific R&D lifecycle. The market is characterized by a shift from simple digital tools to 'agentic' systems capable of autonomous execution. The target audience includes pharmaceutical companies, biotechnology startups, and academic research institutions that are increasingly reliant on computational methods to accelerate discovery and reduce the high costs associated with traditional wet-lab research.

The competitive landscape can be divided into several categories. Direct competitors include other AI-driven R&D platforms and computational biology tools. Established players in the AI drug discovery space, such as Recursion, Insitro, Exscientia, and Insilico Medicine, focus heavily on using ML to identify drug candidates. In the lab informatics and workflow space, Benchling is a dominant provider of cloud-based platforms for life sciences R&D. Additionally, companies like Ginkgo Bioworks and Arzeda focus on synthetic biology and protein design, often utilizing their own proprietary automation and ML stacks.

Synthetic Sciences differentiates itself through its focus on an 'agentic research orchestration layer.' Unlike many competitors that provide specific models or data management tools, Synthetic Sciences offers 'AI co-scientists' that manage the entire research process, from literature review to GPU-accelerated model training and publication drafting. Their emphasis on creating RL environments and capturing process data to enable autonomous systems is a distinct strategic advantage, positioning them as an infrastructure provider for the next generation of AI-led research.

A potential disadvantage for Synthetic Sciences is the presence of well-funded incumbents with deep proprietary datasets and established partnerships with major pharma companies. However, the company's flexible, agent-based approach and its ability to integrate with existing tools (GitHub, Hugging Face, Weights & Biases) may allow it to act as a horizontal orchestration layer that complements, rather than directly replaces, existing specialized tools. As the market trends toward more autonomous and reproducible research, Synthetic Sciences' focus on the 'process' of science could become a key competitive moat.

Founders and Background Report: Synthetic Sciences

Synthetic Sciences was founded in 2025 by Aayam Bansal and Ishaan Gangwani. The founding team possesses a strong technical pedigree in machine learning and computational research, having met while conducting ML research at prestigious institutions including the National University of Singapore (NUS), Carnegie Mellon University (CMU), and the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL). Their collaborative academic work has been recognized in the global research community, with joint publications at major conferences such as NeurIPS, ICML, and AAAI workshops.

Aayam Bansal, Co-Founder, brings prior entrepreneurial and fellowship experience to the venture. He previously founded aisock, which was successfully acquired. Additionally, he has been recognized as a fellow by Z-Fellows and Emergent Ventures, programs known for supporting high-potential technical founders and unconventional ideas. His background suggests a blend of technical execution and the ability to scale a startup from inception to exit.

Ishaan Gangwani, Co-Founder, is characterized by exceptional competitive technical achievements and research depth. His accolades include being a participant in the International Olympiad in Artificial Intelligence (IOAI '25) and achieving USACO Platinum status, the highest tier in the USA Computing Olympiad. Like Bansal, Gangwani is an alumnus of the Z-Fellows and Emergent Ventures programs, highlighting a shared trajectory of elite technical mentorship and support. Together, the founders' expertise in agentic systems and ML research forms the core leadership driving Synthetic Sciences' mission to automate scientific discovery.