Data Scientist
B2B Distribution Analytics
Location
Hours
India - 100% Remote
Structure
7:00am – 4:00pm US Central Time
Reports to
Full-time Contractor
Management Consultant (Engagement Lead)
About the Role
Revenue Optics turns distributor transaction data into selling opportunities. Our clients are US
distributors with 50,000+ SKUs, 100 to 300+ branch locations, and tens of thousands of
business customers. You will take their raw customer transaction data and build the models that
tell a salesperson exactly which customer to call and what to sell them: cross-sell
recommendation algorithms, lapsed product detection, reorder prediction, customer
segmentation, share of wallet estimation, and coverage gap analysis.
Your models do not sit in a notebook. They become dashboards and call lists that inside sales
reps act on the same week, and analysis that goes into steering committee decks reviewed by
CEOs and private equity sponsors. You will see your work move revenue at real companies.
What You Will Own
• Cross-sell recommendation engines: market basket and peer-based recommendation
models across catalogues of 50,000+ SKUs, designed for B2B purchasing behaviour
(repeat industrial buying, not consumer browsing)
• Customer opportunity models: lapsed product detection, reorder prediction, churn risk,
customer lifetime value, and white space scoring across large account bases
• Commercial diagnostics: coverage analysis, sales rep performance quartile analysis,
pricing dispersion, and share of wallet modelling that feed client executive readouts
• Data engineering for messy reality: distributor data comes from aging ERPs with
inconsistent product hierarchies and dirty customer records — you make it usable
• Sales-facing delivery: translate model output into dashboards and account-level
opportunity lists a salesperson can act on without a statistics degree
What Success Looks Like
• 90 days: shipped your first client opportunity model end to end, from raw transaction file
to rep-usable output
• 6 months: cross-sell recommendation methodology documented and reusable across
clients. build once, reuse always
• 12 months: Revenue Optics has a productized analytics engine that you built, deployed
across multiple clients
What We Require (Hard Requirements)
• 5 to 9 years in data science with B2B commercial data: transactions, customers,
products, sales. Consumer recommendation experience alone does not transfer cleanly
to industrial B2B buying patterns
• Production experience with recommendation systems, market basket analysis, or
propensity modelling on large transactional datasets
• Strong Python and SQL, comfort with large messy datasets (millions of transaction lines,
tens of thousands of SKUs)
• ERP-sourced commercial data fluency (must-have): has personally worked with raw
transactional data extracted from an ERP or core commerce system (invoice lines,
product hierarchies, customer master records, location or branch structures) and dealt
with its dirtiness. Candidates with Epicor ecosystem exposure (Prophet 21, Eclipse,
Kinetic) are a premium signal, see the sourcing note below.
• Ability to explain a model to a non-technical sales leader in plain English and defend an
analysis in front of executives
• Sustained US shift commitment: 7am–4pm Central (approx. 5:30pm–2:30am IST)
Strongly Preferred
• Direct exposure to the Epicor ecosystem, especially Prophet 21 or Eclipse (the ERPs
our distributor clients run)
• Experience with distribution, wholesale, industrial, retail, or B2B marketplace data (SKU
catalogues, branch hierarchies, account-based purchasing)
• Recommender systems at B2B commerce companies or SKU-level work at retail
analytics organizations
• Dashboard delivery (Power BI, Tableau, or custom) consumed by commercial teams
• Exposure to ERP-sourced data and its problems (inconsistent hierarchies, duplicate
records, dirty demand signals)
Why This Role
Most data scientists never see their model change what a salesperson does on Monday
morning. Here that is the whole job. Direct exposure to US C-Suite clients, a founder who knows
this data cold from 25 years as an operator, and the chance to build the analytics product of a
fast-scaling firm.