Data Product Architect for BFSI — and a builder shipping real products with AI.
I solve the hard problems inside Finance & Data organizations — the ones where enterprise ERPs meet messy source systems, where reports don't match, where leadership wants insight but the pipelines and architecture won't cooperate. 14+ years, 6 institutions, countless production systems later, I'm now bringing GenAI into those same workflows — and shipping my own products with it on the side.
"Great data products aren't built by adding features — they're built by removing the friction between a leader's question and a trustworthy answer."
The short version of a long career: six deep-dives into data, product, and consulting problems at scale — the business context, what I did, and what changed as a result.
BCG's finance reporting was spread across legacy systems that couldn't keep pace with a global ERP implementation. Corporate reporting cycles were slow, data quality reviews were manual, and leadership needed a credible roadmap for what "Future Finance" actually looked like.
Built and lead a 15-member techno-functional squad. Defined the Finance & Accounting data strategy, secured Investment Committee funding by articulating the business value of each initiative, and embedded directly in ERP implementation planning. Pioneered AI adoption — launched Finance Buddy (a GPT assistant for finance teams) and deployed a Claude-powered data quality validation workflow.
IndusInd needed to launch INDIE — a neo-banking product competitive with fintech-native players. The challenge: ground-up product journeys across savings, payments, investments, and credit, coordinated across Technology, Risk, and Controls teams, against an aggressive launch timeline.
As VP & Product Head – Digital Banking, owned the full product roadmap and launch plan. Managed senior product managers designing customer journeys. Drove cross-functional alignment between engineering, risk, and compliance to resolve the thousand small tradeoffs that decide whether a digital bank ships on time.
A US-based sustainable neo-bank came to Xebia with a specific operational crisis: the call center was their customers' primary channel — and it was failing them. Turnaround times were inconsistent across processes. Agents had no unified view of a customer's profile or history. Ticket statuses weren't getting updated. Different grievance teams operated in silos — so a repeat complainant would reach out and be treated as a new case, each time, by a different agent who had none of the earlier context. For a digital bank whose customers *chose* to call, this was existential.
Led an 8-week pre-sales consulting engagement from Xebia's expert team. Mapped the customer journey across call center operations, re-engineered 8–10 core service ticket processes, and designed a unified omni-channel call center platform with a single customer 360° view at its core. Pitched the solution to the client's leadership with a clear before-and-after view of customer experience and agent productivity.
The pitch was successful. Xebia was awarded the build contract for the unified CRM platform — a direct result of the consulting diagnosis and the re-engineered process blueprint.
DigiBankAsia was building UNO Bank — the Philippines' first full-spectrum digital bank — targeting one of Southeast Asia's most underbanked markets, where roughly 70% of the population had no bank account. They were assembling the launch stack: Mambu for core banking, Backbase for customer experience, AWS for cloud. What they needed from Xebia was a digital engineering partner who could help evaluate the core banking platform choice, shape Day-1 banking services, and make the integration architecture work against a regulatory launch clock.
As Principal Consultant at Xebia, supported Day-1 business advisory for the launch. Evaluated Mambu's feature set against UNO's full-spectrum banking requirements as part of the CBS platform decision. Drove data strategy for the new bank — how customer, transaction, and product data would flow between Mambu, Backbase, and downstream analytics systems. Xebia emerged as UNO's trusted digital engineering partner for the launch.
UNO Bank secured its digital banking licence from the Bangko Sentral ng Pilipinas in June 2021 — becoming one of the first full-spectrum digital banks in the Philippines. The Xebia–UNO partnership was publicly announced in the fintech press.
Yes Bank's frontline — Relationship Managers in branches, service teams on the phones, sales teams in the field — had no unified view of the customer. 30+ source systems, each siloed. A customer walking into a branch and one calling the contact centre the same day would be treated as two unrelated events. RMs were making cold calls based on hunches. Marketing ran campaigns on static segments. The bank's stated ambition — to become "a technology company in the business of banking" — was bottlenecked by a data layer that couldn't keep up.
Architected YESGenie from the ground up — the first-of-its-kind omni-channel (mobile + web) employee super app among Indian banks. Designed the entire data layer: 30+ source systems unified into 20+ data microservices, built on a Cloudera Hadoop stack with Spark and Kafka powering real-time triggers. Led a 50+ member cross-functional team of engineers, modellers, and SMEs.
The point wasn't the plumbing. It was what the plumbing made possible: intelligent, data-backed Next Best Actions delivered to RMs in near real-time. If a customer received a credit, the RM got a notification to pitch an FD. If a customer attempted to prematurely break an FD, the RM was alerted in time to pitch an overdraft against that FD instead. We built a behavioural persona tagging framework using machine learning on transaction patterns — if a customer paid school fees every month, they were tagged "has school-going kids" and surfaced as a candidate for Child Investment Plans. A customer booking a large travel-site transaction was tagged "about to travel" and routed a travel card or travel insurance pitch. An entirely data-backed marketing engine, replacing cold calls with intelligent, context-aware conversations.
YESGenie served 1,000+ branches with a unified platform. Report-to-market time dropped by 50%+. The bank ran hundreds of terabytes of data analysis daily on the Cloudera stack. The work was publicly profiled in the Cloudera customer story and covered by Express Computer as a milestone in Indian banking's data journey.
For its time — 2018–19, a commercial bank running real-time ML-driven Next Best Actions on a Hadoop stack across 30+ systems — this was well ahead of where most Indian banks were. I'm still proud of it.
U Gro Capital's central data warehouse couldn't keep up with growing AUM, regulatory reporting demands, and an LMS migration that touched every core entity. Reports to regulators and the board needed to be bulletproof.
Led the Business Intelligence Unit. Redesigned the central DWH with stronger ingestion controls. Spearheaded the LMS migration — restructuring core data entities so the business could scale and the compliance team could sleep at night.
Borrowing (with respect) from a much greater autobiography. This is the honest log of what I'm learning to build with AI — inside the enterprise, and in my own time. No hype, no demos. Just things that actually ship, and what they taught me.
"My purpose being to give an account of various practical applications of these principles, I have given the chapters I propose to write the title of The Story of My Experiments with Truth." — M. K. Gandhi, 1927
Two production deployments inside Finance Data at BCG. The goal wasn't to look clever with AI — it was to remove friction from workflows that finance teams repeat every week. Both are live, both measurably changed how work gets done.
A GPT-powered assistant for BCG's finance team — helps them answer questions, navigate processes, and surface insights. Built with guardrails, curated training data, and a clear product surface. Not a demo; a daily tool that people keep coming back to.
Replaced a manual, weekly data-quality review cycle with a Claude-powered workflow. The agent reviews anomalies, flags issues, drafts explanations. The team now spends time on decisions, not on looking things up. The pattern I'm most excited to replicate across finance operations.
Three things I built outside of work using Claude Code and Codex. The point of each was to answer a real question: can a senior operator, without an engineering team, ship something polished in a weekend? Turns out — yes.
A full tournament and match-management platform for the pickleball community. OTP authentication, multi-format league scheduling (groups, knockouts, playoffs), live standings, match logging, head-to-head stats, player challenges. Cloud-synced, mobile-first. Built end-to-end in Claude Code — data model to deployment.
A warm, conversion-focused website for my wife Aanchal's tarot practice. Custom visual identity, responsive design, integrated WhatsApp/email booking flow. A whole small business's web presence, shipped in a weekend with Codex.
The site you're on right now. Editorial design, custom typography, case-study architecture, responsive layout. Designed and coded with Claude — concept to published site over a weekend. Proof that senior operators don't need an agency anymore to build something intentional.
I'm currently open to senior full-time roles, fractional consulting engagements, AI build sprints, and advisory calls. Each track is scoped differently — pick what fits your need.
VP / SVP / Director roles in Data Products, Digital Banking, or Finance Transformation at BFSI firms, consultancies, or well-funded fintechs.
5–10 hrs/week engagements for startups and mid-market firms that need senior BFSI and data expertise without a full-time hire. Ideal for pre-launch digital banks, lending fintechs, or data platform rebuilds.
For teams that want to see what's possible before committing to a full AI roadmap. I'll work alongside your team for 1–2 weeks to ship a working internal tool or prototype using Claude Code, GPT, or n8n — so leadership can see what "production-grade AI" actually looks like in your context.
1–2 hour sessions via expert networks or direct booking on Topmate. For leaders who need a second opinion on a specific problem — a data architecture decision, a build-vs-buy question, a BFSI market view, or an AI use case evaluation.
Short essays on building data products in BFSI, putting AI into real workflows, shipping side projects with Claude Code and Codex, and what senior operators can actually do with these tools in 2026. New posts monthly.
I share work-in-progress thinking there before polishing it into essays here. Follow along if you care about data products, BFSI, or what senior operators can actually build with AI.
Follow on LinkedIn ↗Hiring for a senior data or product role? Building a digital bank or fintech and need fractional help? Want an AI build sprint? Or just a 30-min expert call? Reach out directly — I respond to every serious inquiry.