Akash Sharma

Who is Akash Sharma?

‍

Akash Sharma is the co-founder and CEO of Vellum, a development platform for production LLM applications. He is best known for helping teams move from LLM prototypes to reliable, versioned, testable workflows in production. (vellum.ai)

Background and career

‍

Before founding Vellum, Sharma worked at Dover, where he and his co-founders were building generative AI applications and ran into the practical problems that later shaped Vellum’s product, including prompt versioning, quality measurement, and production reliability. Vellum’s public team page lists him as CEO, and TechCrunch reported that he co-founded the company after that experience. (vellum.ai)

He studied at the University of California, Berkeley, according to his public LinkedIn profile. His career path reflects a common founder arc in AI infrastructure: firsthand use of emerging model capabilities, repeated exposure to operational pain, and then building the tooling needed to support production use. (linkedin.com)

Key facts about Akash Sharma include:

1. Current role: Co-founder and CEO of Vellum.
2. Company focus: Production tooling for LLM applications, including prompt workflows and evaluation.
3. Earlier experience: Worked at Dover before founding Vellum.
4. Education: University of California, Berkeley.
5. Public presence: Shares product and AI-building updates on LinkedIn and through Vellum’s official pages.

Notable contributions

‍

1. Founding Vellum: Helped launch Vellum to support teams building with LLMs in production, after seeing the need for better tooling firsthand. (techcrunch.com)
2. Shaping prompt engineering workflows: Vellum’s early product emphasized prompt comparison, testing, and version control, which are core patterns for production LLM development. (techcrunch.com)
3. Scaling customer adoption: TechCrunch reported that Vellum had paying customers and growing revenue early in its lifecycle, reflecting market demand for the category he helped define. (techcrunch.com)
4. Building with product context: Vellum’s team page describes him as someone who works across product and architecture as a single system, which aligns with founder-led infrastructure companies. (vellum.ai)
5. Advocating practical AI tooling: His public writing and interviews focus on moving AI teams past demos and into dependable deployment patterns. (techcrunch.com)

Why they matter in AI today

‍

1. Production focus: His work reflects the shift from “can it work?” to “can it work consistently?” which is now central to AI engineering.
2. Workflow maturity: Teams need versioning, evaluation, and review loops, not just prompt experimentation, and Vellum sits in that gap.
3. Cross-functional adoption: Vellum’s story shows how product, engineering, and operations teams increasingly share ownership of LLM quality.
4. Infrastructure mindset: His approach is useful for builders who want to treat prompts and agents like software assets, not one-off hacks.
5. Shipping faster: The broader lesson is that the best AI teams reduce manual coordination so they can iterate on model behavior with confidence.

Where to follow their work

‍

The primary places to follow Akash Sharma’s work are Vellum’s official site and team page, plus his LinkedIn profile, where he shares product updates and founder commentary. Those are the most reliable public channels for tracking his work in real time. (vellum.ai)

For a broader view of Vellum’s direction, TechCrunch’s coverage of the company’s seed round is a useful snapshot of the product vision he has been building toward. (techcrunch.com)

How PromptLayer connects with Akash Sharma's work

‍

Akash Sharma’s work at Vellum highlights the same core need PromptLayer serves: helping teams manage prompts, test changes, and understand what is happening in production. For teams building LLM products, PromptLayer gives a practical way to bring structure to prompt iteration and observability while keeping engineering workflows moving.

Ready to try it yourself? Sign up for PromptLayer and start managing your prompts in minutes.