Joseph M. Hahn, Ph.D.
Ph.D. scientist with 12+ years architecting and deploying production AI, machine learning, and generative AI for enterprise clients across many industries. Available for select project and fractional engagements, from small shops to the enterprise.
About
I spent eight years at Oracle delivering AI engagements across manufacturing, oil and gas, finance, public sector, real estate, and retail. That work spanned customer discovery, technical design, and POC-to-production delivery. Before Oracle, four years delivering big-data ML systems on Hadoop for enterprise clients. And before that, two decades of peer-reviewed research in computational physics and planetary dynamics.
That combination is the differentiator: I am a scientist who can sit with the C-suite on Monday and write the production code on Tuesday. I work directly with technical and business stakeholders to turn ambiguous AI ambitions into systems that actually run.
As an independent practice, I work across industries and recommend the architecture that actually fits your problem. Vendor-neutral by design.
A small example of the method: this site was designed, built, and deployed with Claude Code in a matter of days. That same compression is what puts a working client prototype in weeks rather than quarters.
Services
What I design and build for you, whether you have a data team or you are the whole team.
Retrieval Augmented Generation (RAG), natural-language-to-SQL (NL2SQL), prompt engineering, and GPU deployment. Each built as the right tool for the specific use case, including which model to run it on: candidates judged blind on identical inputs by a stronger independent grader, so a cheaper model gets chosen on evidence rather than nerve.
Production single-agent, multi-tool systems that combine RAG, NL2SQL, and web search behind one conversational interface. Pointed at a feed of unstructured input such as news, filings, tickets, or intake, the same design makes the routine call a person would otherwise make by hand, with deterministic code handling anything that has to be auditable. Plus emerging multi-agent (agent-to-agent) designs where specialized agents negotiate and coordinate.
Supervised ML on tabular and time-series data: predictive maintenance, risk scoring, forecasting, fraud detection, resource optimization. End-to-end pipelines when you need one, or on-demand ML when you do not: a plain-English question in, a validated forecast out, with no standing deployment to operate.
Working prototypes in weeks, not quarters. AI-assisted development with Claude Code compresses the build cycle, so you see your use case running on your own data before committing to a production budget.
See the Chicago crime forecast demoThe same capabilities above, scoped and priced for a business that does not have a data team.
Fixed scope and fixed fee, agreed before the work starts. If an off-the-shelf tool already does what you need, I will tell you that instead.
What this looks like for a small businessUnderlying stack: Python, SQL, pandas, scikit-learn, LLM APIs, vector search, Spark, Hadoop ecosystem, Jupyter, Git, Claude Code, Anthropic SDK, plus a deep Oracle Cloud Infrastructure stack (Oracle Database, Oracle Autonomous AI Database, OCI Data Science, OCI Generative AI service), as well as conceptual fluency across the major hyperscalers.
Engagements
Two ways in: an independent read on AI work someone else is selling you,
or my hands on the work you are building yourself.
Independent technical judgment on AI work you are buying from someone else. I am paid by you and by nobody selling to you: no vendor relationships, no referral fees.
Read the proposal $2,500 · five business days
Send me the proposal, SOW, or RFP response. You get back whether the architecture, data assumptions, and pricing hold up, what is missing, and what to ask before signing.
Read the opportunity $12,000 · three weeks
Three weeks inside your firm. I talk to the people doing the work, find out what data actually exists, and rank where AI would move the numbers, what each should cost to build and run, and what to skip. The cost line is what you hold a vendor's quote against.
Best for: an AI proposal on the desk, or a pilot about to be funded, and nobody in-house whose job it is to judge it.
$5,000 / month
Flat monthly fee, so the cost is the same whether the pilot runs long or ends early.
Best for: a pilot where the only people with the expertise are the ones being paid to deliver it.
I work directly on your AI projects, hourly or monthly. The retainer rate runs above the project rate on purpose: you are buying reserved capacity, a committed response window, and continuity of context, not just hours.
$250 / hour
Fixed-bid pricing available for well-scoped projects.
Best for: a strategy sprint, a POV build, a custom RAG or agent integration, or any ML or data analysis task.
From $3,500 / month
10 hours at $350 / hour, or 20 hours at $300 / hour. Hours do not roll over. Cancel anytime.
Best for: teams that want a senior AI hand continuously engaged with their stack and roadmap.
$5,000 fixed
One scoped build, priced before it starts, so a business without a budget line for consultants knows the number up front.
Best for: an owner or a small team with real data and one concrete AI job to do, and no interest in an hourly meter.
Custom engagements welcome. Multi-month projects, advisory boards,
and
equity arrangements considered for the right fit.
Selected Work
Predictive ML for industrial process safety
Supervised ML on sensor time-series, alerting operators before dangerous conditions develop on the production floor.
ML for local-government revenue workflows
Supervised regression on tabular records, supporting a local-government revenue-collection workflow.
GenAI workflow automation for a public-sector case-management team
Auto-drafting structured reports from unstructured intake, dramatically reducing analyst time per case.
Natural-language analytics for a specialty e-commerce retailer
NL2SQL self-service analytics that lets non-technical business users query merchandising and sales data in plain English.
Media research AI agent with web search
AI agent with multiple tools that automates ongoing data collection from web sources for downstream analysis.
Capabilities
The working stack I build with, and the judgment I bring when the build is someone else's.
Get in touch
Tell me what you are trying to build or untangle, and I will reply within one business day with a candid read on fit, scope, and next steps. No obligation.
Email Joe