Sign In

Ashish Kakran

VC investor in early-stage Gen AI & cloud infrastructure startups | Helping technical founders build enduring companies | Investor in Harness, Cohere, Isovalent, Opaque..

Ashish Kakran is a venture capital investor with a focus on early-stage cloud, AI/ML, and cybersecurity startups. He works closely with technical founders to help build sustainable companies by providing a blend of technical expertise, customer insights, and empathy.

With a background in software engineering and product management, Ashish brings a wealth of experience to the startups he invests in. He has a deep network of industry professionals that he leverages to support the growth of the companies he works with.

Apart from his professional endeavors, Ashish is passionate about community building and connecting with people. He holds a Bachelor's degree in Computer Science from IIT and an MBA from the Kellogg School of Management.

Some of his recent investments include Cohere, Harness, Isovalent, CyCognito, and Opaque. Ashish has been recognized as a rising star in the venture capital industry by VCJournal and Business Insider.

His specialties include venture capital, product management, CXO networking, investing, deal structuring, inside sales, business development, cloud infrastructure, cybersecurity, data infrastructure, startups, agile software development, DevOps, MLOps, and board responsibilities.

Highlights

Aug 19 · linkedin

I’m excited to share that my latest article on transforming transformers is now live on VentureBeat 🚀! Hundreds of billions of dollars have already been spent on #AI models built on the Transformer architecture. What is the next big thing when it comes to foundation models?

Recent model launches from Cohere, AI21 Labs, Databricks, and SambaNova Systems are making an impact, continuing to shape the rapidly evolving AI landscape. Data/AI leaders seem to have moved from 'one massive model to solve all problems' mindset to smaller model + domain-specific data = better/faster/cheaper performance. Just as individual teams currently select the best-in-class tools for cloud infrastructure, we will likely see different enterprise teams choose the most suitable models for their specific use cases. Already there are so many options available to a team as it rolls out AI products:

💡 open/closed source models now that there seems to be feature parity 💡 large/small language models depending on the use case 💡 state space models to better handle long sequences 💡 Composition of experts / Mixture of Experts for better accuracy 💡 To Fine-tune a model and/or use RAG architecture.

What research areas have captured your interest the most? I’d love to hear your thoughts—feel free to share in the comments or send me a DM. I'm always up for a brainstorming session, especially with founders starting something new!

Thomvest #founders #generativeai

Aug 6 · linkedin

🌟 Congratulations to Harness, Cohere and Carta teams on the recognition among 2024 Forbes Cloud 100 list! At Thomvest (Umesh, Don) we feel honored to be a part of your journey. I am sure your continued success will inspire more technical founders to launch their dreamboats. We will continue to work hard to earn the right to work with them.

#ForbesCloud100 #CloudComputing #AI

Ashish Kakran
Add to my network

Location

San Francisco Bay Area