EY-Parthenon

Revenue Anomaly Detection

Vectorized Python tool that flags eight classes of anomalies across Customer × Product × Month revenue data, including abnormal zeros, negatives, month-on-month spikes, product offsets, and customer churn signals. Outputs flagged pivot grids and summary statistics exportable to Excel and HTML.

Account Deduplication & Entity Resolution

TF-IDF character n-gram similarity with union-find clustering and an LLM-assisted validation step for uncertain pairs. Produces canonical account-name mappings that clean up revenue roll-ups across messy customer lists, a data-quality problem that badly needs solving on real deals.

Contract & Invoice Analyzer

Extracts commercial terms from batches of PDF and Excel contracts: customer, products, amounts, currency, dates, service periods. Flags non-recurring line items, compares clause patterns across a document set, and reconciles against a client account list. Every extracted field is anchored to a source quote for auditability.

ARR Global Enabler

Multi-page branded documentation and onboarding platform for Databricks and Unity Catalog deployment, built for global teams to self-serve the firm's ARR analytics application without a setup call. Covers automated and manual deployment paths, data-cleansing steps, and a VS Code + Databricks workflow.

Buy-side Target Research

Structures public and regulatory data into investment narratives: per-subsidiary recurring versus one-off revenue classification, recurring-revenue share estimates with confidence levels, and a synthesis briefing that feeds downstream ARR modelling.

Team Scaling

Joined as the team's fifth member. As the team scaled past 65 people, onboarding and coaching new joiners became a central part of the role.

Training & Workshops

Designed and delivered two full training programmes to over 50 team members. Travelled to EY GDS India in 2024 to deliver on-site training as a consultant. Ran hands-on workshops for teams across Western Europe.

Ways-of-Working Redesign

Part of the team's migration from Alteryx workflows to Python and SQL on Databricks and Snowflake, with Azure data lakes and a repeatable VS Code-based delivery approach. Built onboarding documentation so the new operating model could be adopted without a one-to-one handhold.