Assumed Normal LLC · Data science & AI advisory · Austin, TX

Put your data, and AI, to work. Without hiring a data team.

I’m Max Mitchell, Head of Data Science at Veritus Agent, a YC-backed AI agent platform. Over 14 years I’ve built data science at Cloudflare, Goldman Sachs, MLB, Divvy Homes, and Upstart. Now I help small and medium-sized businesses get numbers they can trust and adopt AI where it saves real time and money.

Book an intro call See services
14 yrsData science & analytics
2×Head of data science
#1First DS hire, Cloudflare
M.S.Statistics, Carnegie Mellon
The AI on-ramp

A practical path to AI for small and medium businesses.

Most AI projects stall because the data underneath isn’t ready, or nobody picked a problem worth solving. I start with your data and your workflows, pilot one high-value use case, measure it, then scale what works.

  • Automate the weekly report someone builds by hand
  • Sort and route inbound emails, tickets, and requests
  • Pull structured data out of PDFs, invoices, and forms
  • Let your team ask questions of clean, governed data
  • Get your staff productive with AI tools, safely

  1. Assess

    Audit your data, tools, and recurring work. Deliver a ranked map of where AI and automation pay off, and what has to be fixed first.

  2. Pilot

    Build one use case end to end with a clear baseline, so you know whether it saved hours or dollars before spending more.

  3. Roll out

    Harden what worked: reliable data pipelines, simple guardrails, human review where it matters, and training for your team.

  4. Operate

    Ongoing advisory to monitor quality, tune models, and choose the next opportunity as tools change.

Services

What I help with

A monthly advisory retainer for senior judgment and direction, plus fixed-scope projects when something needs to be built. You get the head of data science without the full-time hire.

01

AI & automation

Use-case selection, pilots, and rollout of AI tools that fit how your business already runs.

02

Analytics & reporting

KPIs, one source of truth, and dashboards leadership trusts. No more dueling spreadsheets.

03

Forecasting & models

Demand and cash forecasts, churn, pricing, and lead scoring, built on a solid data foundation.

04

Data engineering

Automated pipelines, a clean warehouse, and integrations so reporting runs without anyone babysitting it.

Services and engagement model →

Selected work

Built and shipped

Production systems, not slide decks. A sample of work from my in-house roles, the same skills I now bring to smaller teams.

Divvy HomesReal estate

AI triage for inbound communication

Shipped a production classifier that decides when an AI-based property management platform needs a human to respond, cutting operational load.

Divvy HomesReal estate

Rent and home value estimators

Built a rent estimator from zero: scraping and public data, feature engineering, deployment, and monitoring. It became the core pricing signal across operations. Paired it with a home value model that learned from operator feedback.

Veritus AgentAI agents · lending

Voice agent performance analysis

Regular analysis of AI voice agents to find weak spots and recommend concrete fixes for conversation quality, compliance, and recovery rates.

Marcus by Goldman SachsConsumer lending

Direct-mail acquisition decisioning

Owned the response models, risk models, and return-optimization framework that decided who received offers on every campaign drop.

MLB Advanced MediaSports

Fan retention and scheduling forecasts

Predicted ticket re-purchase so teams could target at-risk fans, and forecast revenue for candidate schedules ahead of league-wide votes.

CloudflareInfrastructure

Stood up data science from scratch

First data science hire. Built streaming log pipelines in Go, Dataflow, and Spark, and automated revenue tracking by integrating billing and accounting systems.

Fit

A strong fit if you…

  • Run a small or medium-sized business with real operational complexity: professional services, healthcare, trades, e-commerce, lending, real estate
  • Lose leadership time to debating whose numbers are right, or waiting on manual reports
  • Want to use AI but aren’t sure where it would pay off, or whether your data is ready
  • Want senior data leadership without a $200k+ full-time hire
Where I’ve worked
Veritus AgentUpstartDivvy HomesMarcus by Goldman SachsMLB Advanced MediaCloudflareTransUnionU.S. Census Bureau

Let’s find your first win.

A 20-minute call is usually enough to tell whether a diagnostic, a retainer, or a single project makes sense.

Book an intro call