Asher Dean Goldstein

Work and projects

Work

What I've worked on at Simmons Foods, at FFI Solutions and on my own. I keep company details general.

Experience

Projects

Acquisition target screen

Fall 2026

Simmons Foods · Corporate strategy

An always-current, AI-built watch list of 1,000 potential acquisition targets in markets next to Simmons: 12 criteria, AI scraping and Python, built in about a week instead of roughly 1,500 hours.

Read the full write-up →

Claude for supervisors

Fall 2026

Simmons Foods · Operations technology

An AI workflow system supervisors run from their phones, built with Claude skills and Microsoft Power Automate, so they spend less time at a computer and more on the floor. Five supervisors volunteered for the pilot.

Read the full write-up →

Computer vision use cases

Fall 2026

Simmons Foods · Plant and data science

Problem
The plant and the data science and AI team needed to agree on where computer vision would pay off on the floor. My main focus is reducing dead-on-arrival birds, which cost money and reduce yield.
Approach
Most of the work isn't modeling. It's figuring out what the plant isn't measuring yet, building that data collection into workflows people already follow, and only then applying AI to data rich enough to be worth it.
My role
I'd built a working relationship with the data science and AI team, and from the floor of the Benton County plant I worked to bridge the gap between that team and the plant: I helped prioritize which use cases were realistic and which weren't worth pursuing (cameras fog up in some areas, for example), and connected the team with the right people to ask the right questions.
Use cases
  • Flagging dead-on-arrival birds in live receiving to cut costs and protect yield.
  • Spotting where in the process bruising and scalding happen.
  • Measuring how much meat is left on the frame after deboning.
  • Catching repetitive failures like broken belts, so people spend less time on routine checks.
Status
Identifying and scoping use cases.
What I learned
Cross-functional work depends on translation. Plant teams and data teams often describe the same problem in different terms. Much of my value was translating between them, so good ideas didn't get lost and weak ones were dropped early.

AI reserves-data extraction tool

Summer 2026

FFI Solutions

Problem
FFI sells a product that helps investors see their fossil fuel exposure. The research behind it was done by hand, roughly 300 companies at a time, which was slow and capped how far the database could grow.
Approach
Built and tested an AI-assisted workflow that finds companies' annual filings and pulls out proved, probable and possible reserves, handling the different sections, languages and units companies report in. I designed it around the team's existing research process, not technology for its own sake.
Protecting the data
The numbers feed investor-facing analysis, so the tool flags uncertain filings for a person to review instead of guessing a value.
Result
Showed it could grow the firm's reserves database by 80% and cut research time from weeks to minutes.
What I learned
Building a tool is one part of the job. Getting people to buy in and adopt it is another, and often the larger one: change management matters as much as the build.

Contact

Let's talk.

I'm looking to start my career in consulting, ideally working on consumer packaged goods, food, agriculture or operations. I'd love to hear from consultants, recruiters and anyone in that world, in Arkansas, New York City, Dallas or beyond.