Andrew Edmond

Independent buy-side technical due diligence

Southern Spain · Remote
On your deal timeline

§1Summary

Technical diligence for the AI era.

Most engineers who can assess an AI-built codebase have never sat on the buy side. I've done both — five AI-capability acquisitions at Amazon, and production software shipped solo this year.

andrew@andrewedmond.com · Book a call
§2Scope

What I do

Buy-side diligence

Typically two to four weeks, run on your deal timeline. One assessor, close to the code.

  • Architecture
  • SDLC & engineering practice
  • Security posture
  • Org capability & key-person risk
  • AI provenance & maintainability
  • Risk register for the IC

Sell-side readiness

The same assessment, run before a process starts — so the findings are yours to fix, not the buyer's to discount.

Post-close assessment

For portfolio companies, after the deal — where the engineering risk actually gets managed.

§3Method

The Framework

Assessing a codebase the AI wrote.

A public rubric for AI-assisted codebases. Each criterion becomes a rated finding in the register you hand your investment committee.

01 Test coverage relative to generated volumeIs the safety net keeping pace with how fast code was produced?
02 Architectural coherenceOne considered design, or many local answers stitched together?
03 Dependency sprawlWhat was pulled in, and who understands it now?
04 Commit & review patternsDoes the history show engineering judgment, or accepted output?
05 Can the team maintain what the tooling produced?The question the whole assessment resolves to.
Framework — draft in place
Andrew to write the assessment guidance under each criterion. The five rows above are the brief's own enumerated signals; the prose that turns each into a rating is yours to write. The ratings shown are illustrative of the register grammar, not a live assessment.
§4Evidence

Current hands-on practice

Get Camino

A Spain relocation planner — getcamino.app. Empty repo to live product in under a week: ~13,700 lines of TypeScript, five languages, iOS shipped. The build log documents how it was engineered.

SiltHQ

Distributed file infrastructure, in progress — silthq.com.

Source available for the projects above on request, or on GitHub — engineered to the standard the framework describes.

§5Background

Background

Nearly thirty years in engineering leadership — CTO, VP Engineering, and principal-level roles. Principal, M&A Technology and Engineering Integration at Amazon, leading five strategic AI-capability acquisitions for AWS, buy-side. CTO at Woot.com. Engineering leadership at Twitch. AI-first SaaS transformation at Valdera. Computer Engineering, USC. Based in southern Spain; work is remote — deal timelines are the constraint, not geography.

§6Contact

Reach a person

andrew@andrewedmond.com Book a call →

Selective advisory and fractional engagements are taken by referral.