Builder, first.

I'm Toshish Jawale. Before any of the roles (and there have been a few), I'm a software engineer. I like systems: designing them, scaling them, watching them fail in instructive ways, and building the teams that make them reliable.

That started early. I built my first computer at eight, taught myself C at nine, and wrote a snake game the same year. The tinkering never stopped, and it turned into consulting and project work well before I finished my computer engineering degree in Pune. Programming was never a career decision. It was already a habit.

Professionally, I went deep into distributed systems and platform engineering, including seven and a half years at Amdocs. Along the way I started two companies that didn't make it: TimeNome, a time-series database, and UPA, a polyglot persistence abstraction layer. Failed startups are expensive educations, and both taught me things no successful project has. I drifted toward machine learning for a simple reason: the hardest, most interesting systems problems kept moving there. In 2018 that pull turned into a company. I co-founded Symbl.ai with Surbhi Rathore, on the bet that conversations were the most valuable data organizations had and the hardest for machines to understand. As CTO I got to build the whole stack: real-time pipelines that pulled structure out of live speech (work that became a family of US patents), then Nebula, our own language model for conversation, and along the way some published research with collaborators at Georgia Tech.

In 2025 Symbl was acquired by Invoca, where I now lead AI engineering. The work is broader than any one specialty: agents, evaluation, inference infrastructure, building models, and the unglamorous 90% that turns a demo into something enterprises bet revenue on. Running AI engineering at scale is mostly systems thinking and team building, which suits me, because those were always the parts I loved.

Conversation AI remains a recurring territory in my writing (eight years of it leaves a mark), but the through-line is broader: how to engineer AI systems that actually work, and how to lead the people who build them.

Off the clock, I'm usually learning chess (badly, happily; it helps me unwind) or playing the keyboard for an audience of exactly one.

I live in the San Francisco Bay Area. Everything here is my own opinion, not my employer's. The reliable way to reach me is email.