Girish Nadkarni is an Indian-American technology executive and the co-founder and Chief Executive Officer of Together AI, a cloud platform specializing in artificial intelligence infrastructure. Under his leadership, Together AI has become a significant provider of GPU-based cloud services for training and running large language models, competing with established cloud providers by offering flexible, cost-effective access to high-performance AI computing. Nadkarni's work sits at the intersection of Artificial intelligence, Machine learning, and cloud infrastructure, making him a notable figure in the current wave of Generative AI adoption.
Prior to founding Together AI, Nadkarni spent over a decade at Amazon Web Services (Amazon Web Services). As an early member of the AWS engineering team, he contributed to core services that underpinned the first wave of modern cloud computing. He later served as the Head of Engineering for AWS AI and Machine Learning and also led the engineering organization for AWS Trainium, Amazon's custom AI chip designed to optimize cost and performance for training and inference. His experience building enterprise-scale infrastructure deeply informs his approach to making AI more accessible and economical.
Nadkarni co-founded Together AI in 2022 with a team including former AWS colleagues and AI researchers. The company began as a research nonprofit dedicated to the open-source release of large language models and progressed to offer cloud services for model training and inference. Under Nadkarni's executive direction, it became a commercial venture, though it retains a strong commitment to open-source AI, hosting a widely used public platform where models like Llama, Mistral, and various fine-tuned variants can be accessed through an API.
Company Vision and AI Infrastructure
Nadkarni has been a vocal proponent of the idea that the future of AI relies on flexible infrastructure that supports open-source and closed models. He argues that the main obstacles to enterprise AI adoption are not just algorithmic but also technical: high costs, supply-chain constraints for key GPUs like the Nvidia H100, and the complexity of distributed training. Together AI's flagship product addresses this by providing a unified platform with a bare-metal cluster and a model-serving API that abstracts away the underlying hardware (now including AMD GPUs alongside Nvidia's latest models). Nadkarni often frames his company as an alternative to the hyperscaler clouds that also offer AI services (Microsoft Azure, Google Cloud, AWS Trainium), differentiating through a focus on performance for enterprise workloads and transparent per-hour pricing.
Strategic Role in the AI Ecosystem
Nadkarni's influence extends beyond managing a company. He has been involved in several prominent AI safety and governance conversations, participating in forums that discuss the responsible scaling of generative models. His public commentary frequently addresses how to balance efficiency and innovation in frontier AI, particularly the need for sustainable environments that can run large-scale distributed training without extravagant electricity consumption or hardware aging. He has also spoken at major industry events and conferences like Stanford University's AI+ and the All-In Summit, where he shares insights on the practical constraints of deploying AI at scale.
Under his stewardship, Together AI has secured significant funding, including a $200 million Series B round in early 2024, marking one of the largest private investments in the AI infrastructure niche. Nadkarni has been named to various industry impact lists, such as the TechCrunch Disrupt 2024 and the AI 50 list by Fortune and LinkedIn. His perspective is frequently sought by media outlets for commentary on AI industry trends.
Personal Background and Public Representation
While much of his early career is undisclosed, Nadkarni's public profile is centered on his current role. He holds degrees from prominent institutions in his field, and his career trajectory from AWS to founder mirrors the path of several notable AI infrastructure builders. He is an active participant on social media platforms like X (formerly Twitter), where he shares reflections on the practical challenges of large language model-based inference, often referencing Residual Network (ResNet) and Multi-Head Attention developers. He often appears at academic hackathons and developer events, positioning himself as a builder's builder, and engages with the research community to advance the state of efficient AI.
In contrast to leaders at frontier labs like OpenAI or Anthropic, Nadkarni's public role is that of an infrastructure enabler, not a researcher. He emphasizes the deterministic engineering aspects: networking, scheduling, and the kernel-level tuning that make dense computing possible. His prominence aligns with a broader, timely shift in AI towards applications and away from basic research prominence, which he directly contributes to with his\\'s\\' hands-on to powch-lo-based dissemination.
Legacy and Influence
As of early 2025, Nadkarni is a frequent narrator of the modern AI infrastructure arms race. His company is often cited among the valuable players in the "compute layer" of AI, alongside competitors like Groq and SambaNova, which focus on specialized processors and real-time inference. Nadkarni's public statements suggest he views tapping into the immense demand for compute as an opportunity not to hoard, but to democratize. His background in AWS chip design and his current fast-paced startup create a narrative that links the corporate tech giants of the previous decade with the rise of new AI-native infrastructure.
Moreover, Nadkarni is increasingly becoming a connector among the AI elite. He counts collaborations with art and engineering academics and has shares of BAIR (Berkeley AI Research) visiting researcher Marc Tarom. His outcomes in this space foster collaboration between the private and public sectors, and his positioning as a bridge-builder will likely become as significant as his company's -The future: what the next generations of AI infrastructure will be built by those with the depth of experience to meet the brutalities of new tech, a prominent voice in that cohort.