AI/ML Consulting That Turns Data into Working Systems

I help startups and growing companies apply machine learning, optimization, and AI search to real business decisions and workflows. Engagements can range from a focused assessment to building and operationalizing a specific system.

Discuss a project · View selected work


Core Services

Forecasting & Predictive Modeling

Understand what is likely to happen next and where attention will have the greatest value. This work is particularly useful when data is messy, events are rare, or model predictions need to inform operational decisions.

Typical capabilities:

Relevant work: Built a quarterly NYC housing complaint-risk model; approximately 7 million scores were retrospectively evaluated, with 6.3× top-decile lift. Read the case study.

Decision Science & Optimization

Prediction is often only the first step. I help teams turn forecasts and estimated responses into decisions that account for profitability, constraints, and tradeoffs.

Typical capabilities:

Relevant work: Developed pricing and inventory decision systems at Thrasio. An inventory-aware pricing initiative increased profitability by 15% while reducing excess inventory by 20%. Read the case study.

AI Search, RAG & Hybrid Retrieval

Make large, unstructured document collections easier to search and use. The emphasis is on retrieving relevant, traceable source information — not merely generating fluent answers.

Typical capabilities:

Relevant work: Developed and evaluated document retrieval for an approximately 10,000-document corpus, comparing three embedding models and designing a hybrid retrieval approach. Read the case study.

Production ML & AI Systems

Move useful models beyond notebooks and prototypes. I focus on the pipelines, evaluation, monitoring, and feedback loops needed to make an ML system reliable over time.

Typical capabilities:

Relevant experience: At Cadmus, my work includes service-level anomaly detection, monitoring, human feedback, and model retraining workflows.

Technical Audits & Fractional AI/ML Leadership

Get senior technical judgment when a project is ambiguous, a prototype is underperforming, or a team needs help choosing what to build next.

Typical capabilities:

Typical deliverable: A prioritized recommendation, technical assessment, or scoped implementation plan — designed to support a concrete decision.


Engagement Models

Project-Based Build

For a clearly defined forecasting, optimization, AI search, or ML engineering problem. We agree on the use case, success criteria, and deliverables before implementation.

Audit & Advisory Sprint

For teams seeking clarity before committing more time or budget. The result is a focused assessment of data, performance, risks, and recommended next steps.

Fractional AI/ML Lead

For teams that need senior judgment across a roadmap without hiring a full-time ML leader. Support can include architecture, evaluation strategy, reviews, and implementation guidance.

Prototype-to-Production Support

For teams with a promising proof of concept that needs reliable pipelines, integration, monitoring, and maintainability to become useful in practice.


How I Work

My approach is practical and evaluation-driven:

  1. Define the objective: What business decision or workflow should improve?
  2. Assess the data: What information is available, and where are the gaps?
  3. Choose the approach: What level of modeling, optimization, or AI is justified?
  4. Measure performance: What evidence would make the system worth using?
  5. Plan for operation: How will the solution be deployed, monitored, and improved?

The right solution is not always the most complex one. The goal is a system that performs well enough to create value and can be used in the real workflow.


Common Starting Points

You might have valuable data but no clear AI/ML roadmap, an existing model you don’t fully trust, a document collection that’s difficult to search, or a business decision that needs better forecasting or optimization. We can begin by identifying a practical opportunity and the evidence needed to justify investment.

Discuss a project · View selected work


Contact Options