Tristin Esfandiari smiling in a modern office

TristinEsfandiari

Operator, founder, educator

Building the thing. Teaching the room.

Work with me
The record, newest first2026 to present

The operator era

The founder years, compressed into a practice. I get called in when a company has more information than it can use, and the first decision is the one that matters: what to systematize, what to automate, and what to leave to a person. The build comes after, and it is the easier half. M&A, healthcare, staffing, and creative services. I teach at Rady and finish the MBA in between.

M&A advisory firm · AI acceleration seriesRedacted
  1. Market analysis. Comparable sales and active listings for a [redacted] business in [redacted], pulled into one view the advisor can question.

  2. Earnings normalization. Owner add-backs and one-time items reconciled across [redacted] of statements, every adjustment traceable to its line.

  3. Valuation. A draft opinion-of-value range for [redacted], with every assumption cited so the advisor can challenge it.

  4. Deal book. First-pass sections of the confidential information memorandum for [redacted], written to the firm's template.

Client, industry, geography, and deal terms stay inside the engagement.

The work, redacted

Real workflows, with the deals kept off the page

AI adoption at a 40-person M&A advisory firm: which workflows were worth automating (valuations, CIM drafting, market comps) and which were not, then a biweekly series the firm now runs without me. This is the shape of that work, shown to exactly the depth a client NDA allows.

AI adoption, end to end

40-person M&A advisory firm

Scoped which workflows were worth automating (valuations, CIM drafting, market comps) and which were not, then handed over a biweekly series the firm now runs without me. Halfshade came out of it: a redaction layer so associates could push client documents through commercial AI tools without exposing deal data. Confidentiality is the specific reason most advisory firms stall on AI, and this removed it.

Operations that pay for themselves

Healthcare · creative services

At a rehabilitation center, an admissions review that took hours now takes minutes, returning reimbursement, cost, and case-margin estimates without a single insurer call. At an event-content studio, rebuilt offerings and a rebuilt sales pipeline produced $50K+ in new contracts inside 30 days. Both AI workflows are still running without me.

Teaching at Rady

Eight consecutive quarters, lead TA

Lead teaching assistant across eight consecutive quarters, delivering full lectures in Marketing, Business Ethics, and Negotiations. I built the classroom tools that run alongside the courses: ClassCheck for QR attendance and MGT Grader for negotiation roleplay.

Venture work

KRTR.ai × Rady Venture Fellows

Led a six-person Rady Venture Fellows team conducting customer discovery for KRTR.ai: structured interview programs, investor and founder ICPs, and a cold-outreach engine.

Advisor, The Basement

UC San Diego

Advising the startup incubator that launched my own ventures. Full circle.

Rady School of Management MBA, in progress. Sep 2025 to Jun 2027.

2024 to 2026

The founder years

Learned to code in six months, then architected the whole thing.

In February 2024 I founded UniWise, a healthcare AI company building donor-screening and practice-operations software, and ran it through the full venture lifecycle: $450K raised across angel and friends-and-family rounds out of 100+ investor conversations, roughly 1,000 interviews to hire 12 computer-science PhDs and graduate students, and three product phases in twenty-three months. Then I rebuilt the roles around what each person turned out to be strongest at rather than the job descriptions I had written.

I learned to code in six months, then architected the system and trained the team to ship enterprise software in a regulated environment none of them had worked in before. The last product was ContraSight: 300 to 500 page donor records turned into cited, severity-ranked case reviews a clinician could approve, with HIPAA and FDA compliance built in from the first commit — PHI handling, human-review gates, audit citations, and BAAs negotiated directly with covered entities. The final call stayed human by design, which is what made the system defensible to the FDA and acceptable to the clinicians using it. We won first place at UCSD's Basement Demo Day and third at Rady StartR two weeks later, and converted the company to a Public Benefit Corporation.

The UniWise team holding the oversized $5,000 first-place check at UCSD Basement Demo Day 2025
The team with the first-place check, May 22, 2025

First place at UC San Diego Demo Day 2025, the celebration of a decade of student innovation at The Basement.

raised out of 100+ investor conversations
$450K
raised out of 100+ investor conversations
candidates interviewed
~1,000
candidates interviewed
engineers at peak
12
engineers at peak
first services contract
$46K
first services contract
Demo Day and Rady StartR, two weeks apart
1st & 3rd
Demo Day and Rady StartR, two weeks apart

First product · 2024

Collision-repair estimating

Software bridging the two systems every auto-body shop juggles. Shipped, sold against, and set aside when healthcare pulled harder.

Second product · 2024 to 25

Clinic AI for ophthalmology

A $46K contract, then the practice's brand, patient acquisition, EMR, and answering service all launched in 30 days.

Third product · 2025

ContraSight

Cited, severity-ranked donor-record review on Google ADK, the Anthropic SDK, AWS, MongoDB Atlas, and Auth0 — the product this demo preserves.

ContraSight product cover: UniWise branding beside a cascade of product screens for document intake, contraindication review, and case summary

ContraSight, preserved as a working demo

The UniWise platform, unified into one interactive demo on fictional data: follow a single donor from the family screening call, through the AI contraindication review with source-document citations, to the clinical assistant.

Launch the demo

No signup, no backend. It runs entirely in your browser.

MauraLink, live as a working demo

Research to a live platform in six months, with the first enterprise client at 5,000+ employees onboarded pre-launch. Explore it here on fictional data: profiles, pair dynamics, org analytics, hiring, and guided coaching.

Launch the demo (opens in a new tab)

No signup. Fictional organization.

2022 to 2024

The stage and the campus

In 2022 I transferred to UC San Diego and took a full-time job at BAND, the group-communication app run by Naver, South Korea's largest software company. I launched BAND's U.S. education presence — high school bands, dance, cheer, and athletics — running a team of six and splitting the work by what each of them was actually good at, which is how a single user-generated-content campaign pulled more than 10,000 participants.

I also produced five on-location commercials with customer schools, handling the contracts, storyboards, crew, direction, and distribution myself.

On campus, I founded the Student Entrepreneurs Network, UCSD's first student entrepreneurship club: wrote its governance charter, expanded membership to 150, and secured $25K in funding through a pitch to the chancellor. I graduated in December 2024 with a B.A. in International Business, cum laude, on $20K in merit scholarships.

participants in one UGC campaign
10,000+
participants in one UGC campaign
commercials produced on location
5
commercials produced on location
members, SEN at founding peak
150
members, SEN at founding peak
secured in a pitch to the chancellor
$25K
secured in a pitch to the chancellor
2011 to 2022

A decade in marketing before a line of code

2022

Won a public election: Area Representative, Woodland Hills Neighborhood Council.

2022

Founded Repreeve, raised $25K, and used it to buy and forgive roughly $2.5M in delinquent medical debt for strangers.

2019-22

Ran operations across three businesses at Power Washing Pros and OPS Inc — commercial power washing, private security, and virtual staffing — and launched the third one myself, from operating model to contracts to a stack of Salesforce, FieldPulse, QuickBooks, and a custom employee and inventory app.

2019-22

Built the operating infrastructure the business still runs on: website, ad engine, operations software, vehicle tracking. Those systems carried it from four staff to fifty and $8M in annual revenue.

2019-22

Opened a government-contracting channel: cleared the registrations, won public bids, and quoted commercial properties from aerial imagery instead of site visits.

2018

Ran a $1M annual marketing budget at Ivy, a blockchain fintech, through its acquisition.

2017

Founded DAM Inc., my first company: a four-person digital marketing agency with full P&L ownership.

Products, tools, and the code behind them

Thirty-seven repositories, nineteen of them public, all of them shipped for a reason. The ones worth your time:

ContraSight

2025

Cited, severity-ranked review of 300- to 500-page donor records, live interactive demo

Launch demo

Halfshade

2026

Redaction that lets associates run deal documents through AI platforms without exposing client detail

Try it(opens in new tab)

mytxt.dev

2025

The generator: point it at any docs site, get back a standards-compliant llms.txt. 2,800+ sites scanned

Source(opens in new tab)

llmstxt.dev

2026

The directory: a catalog of llms.txt files already live in the wild, built in a weekend

Source(opens in new tab)

MGT Grader

2025

Built for MBA negotiation roleplay: AI grading for recorded negotiation exercises

Source(opens in new tab)

ClassCheck

2025

Built for classroom teaching workflows: QR attendance. React + Supabase

Source(opens in new tab)

OpenClaw

2026

Personal AI agent platform with long-term memory, self-hosted

private

OpenClaw Vault

2026

Turns an Obsidian vault into a shared workspace with your AI agent, synced both ways

Source(opens in new tab)

Autoresearch

2026

AI research automation with experiment memory and an agent-agnostic loop

private

Mauralink KB

2026

The pipeline behind Mauralink: IO-psychology literature in, structured behavioral atoms out

Source(opens in new tab)

Everything else lives on GitHub github.com/tesfandiari1

The AI Pragmatist

A monthly dispatch for executives who have to decide on AI now, not in 2030.

Each issue covers what shipped, what it means for your business, and what to do Monday morning. Written for non-technical leaders who need signal without jargon.

The AI Pragmatist · Monthly

Get the month in AI, minus the hype: what happened, what it means for your business, and what to do about it.

One email a month. Unsubscribe whenever.

I keep room for a few engagements at a time.

I advise M&A, healthcare, staffing, and creative-services firms on where AI belongs in their operations and where it does not. Two shapes: an embedded sprint inside your leadership team (decide which workflows are worth automating and which are not, deploy the ones that are, build the habits that keep them running), or a one-day leadership workshop that ends with your use cases ranked and a 90-day plan. If I'm booked, I'll tell you straight.

  • AI adoption at a 40-person M&A advisory firm, handed off as a biweekly series the firm now runs without me
  • A rehabilitation center's admissions review cut from hours to minutes
  • $50K+ in new contracts inside 30 days for an event-content studio
  • Three years teaching this at UC San Diego's Rady School of Management