AI • Automation • Real Business Outcomes

I build practical AI systems that do useful work.

My AI work focuses on production workflows—not demos. I use ChatGPT and OpenAI APIs alongside conventional software engineering to clean and enrich data, automate repetitive processes, surface customer signals, and help teams make better decisions.

See the Projects Talk About Your AI Idea

SELECTED RESULTS

250,000+inventory records handled by an AI-assisted continuous update workflow

Sales Assistcustomer history turned into next steps, talking points, call scripts, and voicemail scripts

Source-Groundedproduct enrichment constrained to manufacturer-authorized sources, with explicit “Not Found” handling instead of guessing

Applied AI Projects

Examples of how I combine AI with databases, APIs, automation, and business rules to solve concrete operational problems.

PRODUCT DATA

AI Product Data Enrichment

Built a controlled enrichment workflow that researches manufacturer-authorized sources and identifies product line, product name, category, and concise product descriptions. The workflow uses strict output rules, rejects unsupported guesses, and returns “Not Found” when an official match cannot be verified.

✓Official manufacturer sources only

✓Structured machine-readable output

✓Designed for high-volume catalog cleanup

SALES INTELLIGENCE

AI Sales-Assist Agent

Created a sales-assist workflow that combines open quotes, prior orders, customer communications, and internal notes. For each opportunity, AI produces a recommended next step, talking points, a live-call script, and a voicemail script while giving extra weight to explicit customer signals.

✓Uses multiple business-data sources

✓Turns history into actionable outreach

✓Built for repeatable sales-team use

AUTOMATION

250K+ Inventory Update Workflow

Built an AI-assisted automation process to continuously update a catalog containing more than 250,000 inventory records. The goal is practical scale: reduce repetitive manual research while preserving rules, traceability, and human review where confidence matters.

✓Continuous processing

✓Large catalog volume

✓AI integrated into an existing software stack

How I Use AI

I treat AI as another engineering component: powerful when the task is bounded, the data is grounded, and the output is validated.

1. Define the business resultStart with a measurable problem rather than “add AI” as a goal.

2. Ground the modelGive it the right documents, records, APIs, and explicit source rules.

3. Constrain the outputUse structured formats, validation, confidence rules, and clear failure states.

4. Integrate & improveConnect the model to the production workflow, measure results, and iterate.

WHAT AI IS GOOD AT

High-leverage business tasks

Data enrichment, document analysis, customer-history synthesis, structured text generation, classification, workflow triage, and repetitive research where rules can be clearly defined.

WHAT I ADD

Engineering around the model

The useful part is rarely just the prompt. I build the surrounding application logic, database integration, validation, error handling, scheduling, user interface, and operational safeguards that make AI usable in the real world.

Have a repetitive process that should be smarter?

If your team spends time researching, copying, categorizing, summarizing, or deciding what to do next, there may be a practical AI workflow worth building.

Tell Me About It

ChatGPT and OpenAI APIs are tools I use in selected projects. This site is not affiliated with or endorsed by OpenAI.