MLOps podcast #195 with Varun Mohan, CEO of Codeium, Building the Future of AI in Software Development brought to us by QuantumBlack.
// Abstract
This brief overview traces the evolution of Exafunction and Codeium, highlighting the strategic transition. It explores the inception of Codeium's key features, offering insights into the thoughtful design process. This emphasizes the company's forward-looking approach to preparing for a rapidly advancing technological landscape. Additionally, it touches upon developing essential MLOps systems, showcasing the commitment to maintaining rigor and efficiency in the face of evolving challenges.
// Bio
Varun Mohan developed a knack for programming in high school where he actively participated in various competitions. This passion for programming was shared with his now co-founder, with whom he frequently competed. Their common interest in programming and competition led them to attend MIT together, where they undertook more programming challenges. After college, they ventured into the Bay Area where they continued to compete and further cultivate their programming abilities.
// MLOps Jobs board
https://mlops.pallet.xyz/jobs
// MLOps Swag/Merch
https://mlops-community.myshopify.com/
// Related Links
Websites: codeium.com, https://exafunction.com/
--------------- ✌️Connect With Us ✌️ -------------
Join our slack community: https://go.mlops.community/slack
Follow us on Twitter: @mlopscommunity
Sign up for the next meetup: https://go.mlops.community/register
Catch all episodes, blogs, newsletters, and more: https://mlops.community/
Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/
Connect with Varun on Twitter: https://www.linkedin.com/in/varunkmohan/
Timestamps:
[00:00] Varun's preferred coffee
[00:15] Takeaways
[02:50] Please like, share, and subscribe to our MLOps channels!
[03:05] QuantumBlack ad by Nayur Khan
[05:51] Varun's background in tech
[10:55] Language Models Advancement
[14:17] GPU scarce world
[18:23] Vision and Pain Points
[19:18] Fine-tuning Challenges in NLP
[21:04] ML and AI Caution
[21:49] MLOps: App vs Infra
[23:53] Data Engineering Abstraction Evolution
[26:12] Codeium and Scaling Discussion
[31:59] API, Cloud, Computation
[34:20] Codeium scaling
[35:11] Reserved GPUs, companies self-hosting products
[38:00] Open-source code Codeium training
[40:03] Protecting IP Licenses
[41:32] ML Challenges: Data, Bias, Security
[44:37] Evaluating code
[48:29] Getting values from Codeium
[49:49] Exafunction ML AI Production
[52:17] AWS Creation
[53:58] Feature flags and MA AI lifecycle
[56:34] Coding problem
[58:40] New software architectures
[1:03:28] Wrap up
