One dataset.
Two stories.
What does New York’s taxi data leave out?
01 / Independent work & ideas
Independent investigations, working code, and ideas from the classroom.
What does New York’s taxi data leave out?
When does another agent earn its place?
A city, a concert, and a question of attribution.
A small harness for asking harder questions.
Bringing practical AI decisions into the classroom.
01 / 05 One Dataset, Two Stories






02 / My story
From engineering at SRM to two master’s degrees at Penn State. A path through building, learning, and sharing the work.
Master’s degree
Master’s degree
2014 — 2018Bachelor of Technology
SRM Institute of Science and Technology
AI, machine learning, and project delivery.

Project Management Institute
Feb 2025 — Feb 2028

Amazon Web Services
Jan 2025 — Jan 2028
Amazon Web Services
Dec 2024 — Jan 2028

Process improvement

Data analysis & visualization
Sharing the work
Speaking & teaching
Where project management meets AI. A practical session on bringing AI into professional work, for engineers and the people building alongside them.
Guest speakerProject Management Club · UT Dallas
Women Engineers Code Conference
From the personal archive
University life, public conversations, and the communities that shaped my path.
Explore a moment

Giving Tuesday · 2021
Interview clip

Management & Organizational Leadership / Business Analytics

Admissions campaign · 2017

Organizing team

Speaking at “India Welcomes Its Overseas Community”

Attendee
03 / Experience
From business analytics to production AI. Building systems, understanding their failures, and taking responsibility for delivery.
Read my résuméI design and migrate conversational and multi-agent systems for banking, healthcare, and pharma, connecting agent architecture and guardrails with evaluation and client-facing delivery.
My work includes deterministic provenance: audit trails that do not depend on a model’s explanation of its own behavior.
Production agents Regulated industries

Built Python pipelines for conversational anomaly detection across 50,000+ daily messages, with latency under 200 milliseconds.
Developed synthetic datasets and conducted more than 200 fine-tuning experiments, alongside work on AI safety and fairness in healthcare.
Production Python Model evaluation

Promoted from Business Insights Analyst
Joined in business insights and moved into the GenAI team within the COO organization.
Developed automated quality assurance for call-review workflows and worked through six rounds of user acceptance testing to refine system behavior.
Business insights Enterprise GenAI

Supported a B2B/B2C platform migration through business analytics and Power BI reporting.
Business analytics Power BI
04 / Contact
About AI, an interesting problem,
or what we could build together.