About
I build data products and the tooling that makes data teams faster. 8+ years across fintech and e-commerce — shipping production ML, owning company-wide experimentation frameworks, and leading post-merger data-model consolidation at scale. 🔹 Currently at Printify (now FYUL, post Printify + Printful merger) — sole developer of the Retail Price Recommendation system serving 270K+ merchants, and led post-merger unification of Printify and Printful data models (reconciling divergent GMV definitions and designing the merged dimensional layer that powers C-suite dashboards). 🔹 Most excited right now about agentic AI for data teams. Building Claude-powered skills, plugins, and MCP servers that automate the drudge work analysts hate — SQL generation, Looker dashboard scaffolding, LookML linting, automated PR generation. The piece I'm most proud of is a full Power BI → Looker migration agent that parses DAX, data models, and visuals and auto-generates equivalent LookML views/explores/dashboards with validation checks — turning a months-long manual rewrite into an automated pipeline. 🔹 Earlier at Better Mortgage (New York & Gurugram) — operations + product analytics, regression models for loan funding cycle time, vendor performance frameworks. Helped scale the India analytics team from 4 to 20+. 🔹 Represented Printify at the Data Innovation Summit 2023 in Stockholm. Travelled to 15+ countries; worked on the ground in India, the USA, and Latvia. Reach out: ankeshkarn@gmail.com Stack: SQL • Python • Snowflake • Looker/LookML • dbt • Airflow • Power BI • Claude Agent SDK • MCP • XGBoost • Docker/K8s
Work experience
Data Science & AI at FYUL, Dec 2022 - Present · 3 yrs 9 mos, Riga, Latvia, • Agentic AI for data teams (current focus): Claude-powered skills, plugins, and MCP servers that automate Looker dashboard generation, SQL authoring/formatting, automated PR creation, and LookML refactoring — cutting repetitive analyst work by 40–60% per workflow. • Power BI → Looker migration (end-to-end automation): Designed and shipped an agentic pipeline that parses Power BI artifacts (DAX, data models, visuals) and auto-generates equivalent LookML views, explores, and dashboards with validation checks — turning a months-long migration into an automated pipeline. • Snowflake Intelligence agents: Custom agents over Snowflake Intelligence (semantic views + Cortex) serving natural-language analytics to non-technical stakeholders. • Retail Price Recommendation system (sole developer): End-to-end production price-rec engine serving margin-safe low/median/high prices to 270K+ merchants. Hierarchical fallback logic (group → blueprint → variant with branding/channel drop-downs), margin-guard, psychological rounding; deployed via Docker + Jenkins + Helm on Kubernetes. • Post-merger data-model consolidation (Printify + Printful → FYUL): Led unification of two independent data ecosystems — reconciled divergent GMV/revenue definitions, mapped overlapping entities (orders, merchants, catalogs), and designed the merged dimensional layer powering C-suite KPI dashboards. • Experimentation leadership: Drove company-wide democratization of A/B testing; enabled 5+ cross-functional teams to run experiments independently, reducing reliance on central analytics by ~60%. • Causal inference: Synthetic control + DiD evaluation for a cashback campaign after control contamination; informed $M-level marketing allocation. • Survival analysis: Kaplan-Meier framework for enterprise distributor and store activation funnels. • Production ML: Order-processing-time estimation for fulfilment routing; co-owned the Order Routing production-speed model (XGBoost).
Analytics Strategy at Better, Apr 2022 - Dec 2022 · 9 mos, New York City Metropolitan Area, • Owned analytics for volume-management framework redesign to improve unit economics. • Reported business and initiative metrics to VP/SVP operations leadership in weekly executive reviews. • Built regression models explaining drivers of loan funding cycle time (R² = 0.82). • Prototyped cost-prediction models for property valuation. • Developed API integrations across internal/external systems for automated data pipelines.
Senior Data Analyst at Better, Apr 2021 - Apr 2022 · 1 yr 1 mo, Gurugram, Haryana, India, • Partnered with Product and Ops on interventions that reduced loan funding cycle time by ~10 days. • Built monitoring workflows and dashboards that reduced operational backlogs by ~50%. • Designed vendor performance metrics supporting expansion to 3 additional appraisal vendors. • Built performance dashboards tracking productivity, efficiency, and quality for 3,000+ contributors.
Data Analyst 2 at Better, Oct 2020 - Apr 2021 · 7 mos, Gurugram, Haryana, India, • Introduced performance dashboards across 6 appraisal vendors. • Automated 100+ SQL workflows in Metabase/Looker. • Interviewed 50+ candidates; delivered SQL/Looker/Tableau training; helped scale India analytics team from 4 to 20+.
Analyst at LatentView Analytics, Jul 2018 - Nov 2019 · 1 yr 5 mos, Greater Chennai Area, • Lead-generation recommendation engine using RFM scoring and similarity indexing — improved campaign engagement by 6% and drove £5M in additional investments for a financial-services client. • Churn / disinvestment prediction classification model using investor interactions, transactions, consultant ratings, and product performance features. • Power BI dashboard tracking money flows (top-ups, cross-sells, withdrawals) for executive KPI reporting.
Campaign Coordinator at Coalition of Leaders for Sustainable Development, Jul 2017 - Dec 2017 · 6 mos, India
Marketing Manager at Prodigy NIT Trichy, Nov 2016 - Sep 2017 · 11 mos, NIT Trichy, Prodigy is an annual technical symposium of the Production engineering department. As a marketing manager, I will be working in a team to bring down sponsors for Prodigy'17.
Profile highlights
what i've built with AI: 1. 0-1 Agentic Analytics Stack 2. AI experimentation Platform 3. Proactive BI Listening Agent 4. Synergies AI
I'm the person in every team who…: gets most of the complex problem statements
My hot take about working with AI is…: so much fun experimenting
My biggest achievement in college was…: Contributing to development of bullet proof jackets for Indian Army
My proudest non-work achievement: Hopping by the Bangalore lakes!
The role I'd take a 20% pay cut for: AI engineer
