A business executive, entrepreneur and a team builder, Amit has been managing business operations and fund raising, while steering technology companies into global markets. As an OpenSource enthusiast with deep domain expertise in the DevOps and DevSecOps segments, he enjoys the challenge of bringing a fresh and unique approach to solve market inefficiencies at scale. He is the CEO of Kubiya.ai, a platform that leverages conversational AI to turn complex operations into simple conversations – powering the world of AI DevOps Workflow Automation. In an interview, he speaks on an array of topics. Excerpts:

Conception of Kubiya.ai

Kubiya provides a unique ‘embedded’ AI experience for devops and end-users alike. On one hand, Conversational AI chatbot that lives within Slack, Microsoft Teams and a CLI offers a ChatGPT-like experience for organizational users (ex software developers, business users, etc) who are looking to self-service themselves to operational workflows in a simple and secure way. The other side of the coin are operators who through the use of GenerativeAI in a low-code editor can prompt any operational workflow and have it created, validated and tested within seconds. Says Amit, the world of Self-Service for DevOps platforms has largely failed on delivering on the promise of self-service primarily due to the lack of end-user experience. This causes a barrier to adopt, implement, and maintain such platforms which is precisely what the approach Kubiya.ai aims to solve.

Generative AI for Cloud Operations

“We are creating a new approach to DevOps we call Intent-Based-Operations. With all the attention given to Generative AI technologies, we are drawing a clear path to those capabilities in the world of cloud operations.. Effectively our virtual assistant has the ability to turn complex operations into simple conversations by contextualizing the human intent using natural language, filling in all the missing parameters (or having a closed loop conversation to clarify intents), and then serving the user with exactly what they require.”

AI technology- perception vs reality

“AI should not be seen as a license to kill, nor should it be viewed as a replacement for humans.  At the end of the day AI is an enabler for human productivity– albeit a very significant one- that alleviates the heavy dependency on humans in the loop at all times.”   As a technology enthusiast Amit contends that in a perfect world, the human to machine interaction should have clear guardrails and well-defined hand off in place for this to work.  “There is a clear difference between what technology can do, vs what humans will allow it to do. Case in point, technologically it’s already feasible, but how close are we really to a world of self-driving cars?”

Kubiya’s AI DevOps Workflow Automation

One of the defining points is to enable people to turn complex operations into simple conversations using natural language- which makes it easy for any end-user to consume.  “Secondly, we have created a bi-direcitonal feedback loop using reinforcement learning with human feedback (RLHF), a method used to train LLMs on domain specific knowledge which is core to ‘embedding’.   This is similar to the approach Waze took in the world of navigation systems – we allow the end-nodes to provide the system input. Im Kubiya a classic example of provisioning a new environment – a Developer prompts Kubiya: “I want a new development environment in AWS US East 1 Region, with Machine M4 2XL, with security group CG-10231456, and machine imagine AMI-120356ff347”. The system is smart enough to auto-fill the parameters based on the prompts and provision the new environment without needing to go through any additional steps.   The added benefit to this approach is that it personalizes the experience for the users, while removing the heavy lifting from the operators who traditionally would have had to maintain the system on their own.  Now the maintenance of the system becomes democratized across many users with LLMs constantly training on user-interactions which significantly reduces the overhead and cost of adoption.

Leadership traits

Hunger to succeed is one of the main traits Amit exhibits. “Everyone in our company is not only super talented but have an intense hunger to succeed. We check our egos at the door and go to work every single day. Yes, I am very much all the above, but above all I pride myself in being a decent human being who is driven to success for all the right reasons. It’s important to me to help leave this world a better place than when I arrived. As someone who has failed plenty of times in business, I can honestly say I have no regrets and learned plenty from those experiences. However, I’m less afraid of trying something and failing then not have tried it at all.”

“I think as a founder of an AI company, with all the hype surrounding our space, that we need to be conscious not to get drunk off of our own cool aid.  There are plenty of groundbreaking things we can do with AI, but not all of them have a place in business.  Be sure to focus on a real problem and having the AI as an enabler to solve it, rather than building an AI solution and then looking for a problem to solve.”

The Success Mantra

“Keep innovating and always stay hungry.   As someone who has seen firsthand companies with strong success trajectories tail off over time because they forgot how they reached the peak, we need to remember that complacency is our biggest enemy. So celebrate milestones (very important), be thankful for what you have accomplished, reflect on the journey, but always stay hungry.”