A legal document analyser (OCR and retrieval)
BeforeReviewing title documents took 4 to 8 hours each
AfterUnder 30 minutes each
What I build
AI products that turn messy, high-stakes business problems into measurable outcomes, and a way of working that makes sure people use them.
I build AI systems for a living, and I treat getting people to use them as part of the build. At DESRI I started the AI function from scratch and took firm-wide adoption from under 5% to near-universal daily use in under a year, mostly by building what the business needed instead of buying what was on offer.
The climb
Four steps in five years, each with a bigger remit than the last.
2021 to 2023
ExxonMobil
Technical Data Analytics Lead, Bangalore
Built Python and Power BI pipelines for global engineering units, cutting simulation analysis time by 98%.
2023 to 2025
D.E. Shaw & Co.
Senior Analyst, AI and Technology, Hyderabad
Started the AI and innovation function from zero, and took firm-wide adoption from under 5% to near-universal daily use in under a year.
2025
DESRI
Senior Analyst, AI and Technology, Hyderabad
Led delivery of production AI tools across the firm, ran discovery sessions with every function, and two firm-wide AI Weeks.
2026 to now
DESRI
Manager, GenAI Enablement and AI Tech, Hyderabad
Leading enablement and strategy for AI across the firm, and a member of its AI strategy group.
- Under 5% to near-universaldaily AI use across the firm, in under a year
- 4 to 8 hours to 30 minutesto review a legal document, down from a day's work
- 2firm-wide AI Weeks, multi-day design thinking programmes
- 4 of 4interns I mentored, all converted to full-time
- 98%less simulation analysis time, in my first role
How I work
- 01
Discover
Structured sessions with legal, finance, HR and executive teams. Map the workflow, find the pain, and turn what people say into requirements and user stories.
- 02
Co-create
Ideas come from the people doing the work. Two firm-wide AI Weeks, multi-day design thinking programmes, fed the product plan directly.
- 03
Prototype
A working demo before a full build, using fast AI-assisted prototyping. It cuts rework and shows early whether anyone will use the thing.
- 04
Ship and govern
PRDs, agile delivery, and production tools adopted across the whole firm, with governance and enablement so they keep getting used.
What it has produced
A spend dashboard for AI and cloud tools
BeforeCosts scattered across more than ten tools and providers
AfterOne automated view, with reconciliation, per-user attribution and forecasting
A software licence review
BeforePaying for more AI seats than people used
AfterSpend right-sized to actual usage
An executive summary engine
Before300+ page financial and technical reports
AfterReview volume down by 67%
An executive email and calendar digest
BeforeOverloaded inboxes and calendars
After100+ admin hours a month saved
An internal AI workspace
BeforeFragmented tools and manual reporting
AfterUsed across most of the firm, with 90% less manual processing in core reporting
A deal-document assistant
BeforeAnalysts triaging deal documents by hand
AfterA projected 40+ analyst-hours a week back
Firm-wide adoption
BeforeUnder 5% usage
AfterNear-universal daily use within a year
Outside work
I am building personal finance tools for Indian salaried professionals, applying the same loop of research, prototype and ship. Two are live: the True Surplus Calculator and the Salary Decoder.
Education: B.Tech in Petroleum Engineering, MIT World Peace University (2021), and a General Management Programme at Stoa School (2023). The formal record is on LinkedIn.
Building something where adoption matters?
I am happy to talk about AI enablement, product delivery, or a role. Tell me a little about it.