AI Engagement Leader · Ph.D. Linguist
Priyanka Biswas
I take enterprise AI programs from executive discovery through architecture, deployment, and adoption, and make sure people actually use what we build.
AI Engagement Leader · Ph.D. Linguist
I take enterprise AI programs from executive discovery through architecture, deployment, and adoption, and make sure people actually use what we build.
I'm the AI Engagement Lead at KAPI Advisors, where I lead enterprise AI engagements end to end: discovery, solution shaping, delivery planning, UAT, and adoption. I'm the primary point of accountability to client executive sponsors, and I coordinate distributed engineering teams across the US and India to deliver AI agents, digital twins, and enterprise GenAI solutions.
I'm technically fluent enough to challenge enterprise architects on RAG design, agent orchestration, and conversational AI, and I'm just as focused on translating those decisions into business outcomes. I also build the reusable assets (UAT strategies, governance playbooks, executive reporting) that make every engagement start faster than the last.
Before KAPI, I spent over four years at Espressive, most recently as Director of Product Operations (ML), where I scaled the organization from 2 to 14 and owned delivery of enterprise AI agent programs. Earlier, I worked on NLU and speech systems at Google and Apple, and on corpus research at Microsoft Research India.
My foundation is language. I hold a Ph.D. in Linguistics from the University of Southern California and taught NLP-focused courses at UC San Diego. That training still shapes how I think about AI: what people mean, how systems misread them, and how to design for the gap.
AI agents, RAG systems, conversational AI, and digital twins, delivered on platforms like ServiceNow and Microsoft Teams.
Steering committees, delivery governance, risk and dependency management, and executive reporting that keep programs on track.
UAT strategy, change management, and feedback loops so AI lands with the people it's built for and delivers measurable value.
An automated daily briefing that turns the day's AI newsletters into a ranked reading list with time estimates, a recall quiz, and a build idea. It's my own system for staying current without drowning in feeds, and a working example of the kind of AI automation I design for clients.
Read today's brief →Two cooperating agents built with Phidata: a Google Search agent researches and summarizes characters from my daughter's favorite book series, then hands off to a DALL-E image agent that illustrates them.
See the post → AI agent · 2025An Agno agent running Llama 3.3 70B that pulls a video's captions and metadata from a URL and summarizes it, with customized toolkit instructions and a candid write-up of its limits.
Watch the demo →Leads enterprise AI engagements from discovery to adoption and owns the engagement operating rhythm for client executive sponsors.
Owned delivery of enterprise AI agent programs. Scaled the team from 2 to 14 across conversational AI analysis, BI, and dialog content.
Established MLOps practices for model deployment and evaluation, and built tooling that automated annotation and quality checks.
Associate Linguist at Google, improving NLU classifiers through semantic enrichment and data curation. Computational Linguist at Apple, driving a 20% improvement in ASR accuracy.
Nine peer-reviewed publications on quantification, ambiguity, and reference, using formal models and psycholinguistic experiments. Guest faculty for Conversational AI Applications in Business at IIM Kozhikode (2024). Read about my research.
A black belt with international competition medals and a volunteer instructor. The dojo taught me discipline, resilience, and how to stay calm under pressure.
My handmade jewelry business, where I design pieces and run the operations.
Have an AI program to discuss, a speaking invitation, or a question about my work? Send a note and I'll get back to you.
Thanks, your message is on its way. I'll be in touch soon.