ML Engineer
Builds machine learning models that power revenue predictions — from churn scoring and lead ranking to pipeline forecasting and pricing optimisation.
Common Tools
The tech stack for this discipline
MLflow
ML/AIAn open-source MLOps platform for managing the machine learning lifecycle — including experiment tracking, model versioning, deployment, and a model registry — across any ML framework.
Used for
RevOps data science teams use MLflow to track experiments for churn prediction, lead scoring, and deal win probability models — versioning features and hyperparameters so models can be reproduced, compared, and promoted to production.
Python
OtherThe dominant general-purpose programming language used in data science, automation, and analytics, with a rich ecosystem of libraries (pandas, scikit-learn, SQLAlchemy) essential for advanced RevOps work.
Used for
RevOps practitioners use Python to build data pipelines, automate reporting, clean and transform CRM data at scale, build predictive models for churn and lead scoring, and create custom integrations between revenue systems.
Databricks
Data WarehouseA unified data intelligence platform combining Apache Spark-based data engineering, Delta Lake storage, ML experimentation, and SQL analytics on a lakehouse architecture across AWS, Azure, and GCP.
Used for
RevOps and data engineering teams use Databricks for large-scale revenue data transformation, feature engineering for ML models (churn prediction, lead scoring), and building a lakehouse that serves both analytics and machine learning workloads.
Snowflake
Data WarehouseA cloud-native data warehouse platform that separates storage from compute, enabling organizations to store and query massive volumes of structured and semi-structured data across multiple cloud providers.
Used for
RevOps teams use Snowflake as the central analytical layer where CRM, marketing, product, and financial data converge, enabling cross-functional revenue analytics that no single source system can provide.
BigQuery
Data WarehouseGoogle Cloud's serverless, highly-scalable data warehouse that uses SQL and supports petabyte-scale analytics with built-in machine learning capabilities via BigQuery ML.
Used for
RevOps teams at Google Cloud-native organizations use BigQuery to warehouse event streams, marketing touchpoints, and CRM exports for large-scale pipeline and attribution analysis without managing infrastructure.
dbt
ETL/ELTAn open-source transformation framework that enables analysts and engineers to define, test, and document data models using SQL within a version-controlled workflow.
Used for
RevOps and data teams use dbt to build clean, tested, documented transformation layers in the warehouse — converting raw CRM, marketing, and product data into reliable business metrics that power dashboards and models.
Segment
OtherTwilio Segment is a customer data platform (CDP) that collects first-party behavioural data from web, mobile, and server sources, standardises it into a common schema, and routes it to hundreds of downstream tools.
Used for
RevOps teams use Segment to create a single source of truth for customer behavioural data, ensuring the same user actions flow consistently into Salesforce, Marketo, Amplitude, and Intercom without duplicate tracking code.
Monte Carlo
Data QualityA data observability platform that monitors data pipelines and warehouses for anomalies — freshness issues, schema changes, volume drops, and distribution shifts — alerting teams before bad data reaches dashboards.
Used for
RevOps and data engineering teams use Monte Carlo to monitor their revenue data pipelines for failures or anomalies, ensuring that CRM syncs, attribution models, and forecast dashboards are always powered by reliable data.
Clari
Revenue IntelligenceA revenue operations platform that uses AI to aggregate CRM and activity data, forecast pipeline with greater accuracy, and provide real-time deal inspection for sales leaders.
Used for
RevOps teams use Clari to replace subjective rep-submitted forecasts with AI-driven commit and best-case predictions, and to identify at-risk deals through engagement signals and historical patterns.
6sense
AttributionAn account-based marketing and revenue AI platform that uses intent data, predictive analytics, and AI to identify in-market accounts, prioritise pipeline, and orchestrate personalised outreach.
Used for
RevOps teams use 6sense to enrich Salesforce accounts with buying stage predictions and intent signals, enabling sales and marketing to focus outreach on accounts most likely to enter an active buying cycle.
Core Skills
What practitioners do day-to-day
Machine Learning for Revenue
TechnicalThe ability to build, train, and deploy predictive machine learning models for revenue use cases — including churn prediction, lead scoring, deal win probability, and LTV estimation.
Business outcome
ML-powered revenue models allow teams to prioritise the right accounts and opportunities with statistical confidence — improving conversion rates, reducing churn, and allocating rep effort more efficiently.
Feature Engineering
TechnicalThe process of transforming raw business data into informative input variables (features) that improve the predictive power of machine learning models for revenue outcomes.
Business outcome
Good feature engineering dramatically improves model accuracy — turning raw CRM and product data into the signals that reliably predict which customers churn and which leads convert, making revenue models genuinely actionable.
Prompt Engineering
TechnicalThe skill of crafting, refining, and optimising text prompts to elicit useful, consistent, and accurate outputs from large language model AI systems for business and revenue operations tasks.
Business outcome
Effective prompt engineering reduces manual data entry for CRM, accelerates content creation for sales, and enables RevOps to deploy AI automation that is accurate and consistent enough to act on without human review.
SQL
TechnicalProficiency in Structured Query Language for querying, transforming, and analysing data stored in relational databases and cloud warehouses — the most fundamental analytical skill in RevOps.
Business outcome
SQL fluency allows RevOps practitioners to answer any data question independently — without depending on a data team — enabling faster, more accurate decision-making across all revenue functions.
Data Pipeline Design
TechnicalThe ability to architect and build reliable data movement and transformation workflows that extract data from source systems, load it into a warehouse, and transform it into analytics-ready models.
Business outcome
Reliable data pipelines ensure all revenue dashboards reflect accurate, timely data — enabling confident forecasting, attribution, and performance management decisions.
Data Modelling
TechnicalThe ability to design logical representations of business data — defining entities, relationships, grain, and hierarchies — that accurately represent revenue processes and support reliable analytics.
Business outcome
Well-designed data models ensure metrics are calculated consistently, data is non-duplicated, and analytical queries run accurately — giving every team a shared, trustworthy view of revenue.