Anthropologist of technology · Mixed-methods research · AI evaluation & governance

Innovation doesn't start with technology. It starts with people.

I'm Shaheen Amir, an anthropologist of technology. For fifteen years one question has held steady — how do people actually live with the systems we build for them? — and it has only grown more urgent as we deploy models faster than we can examine them. Bridging qualitative and quantitative research across Microsoft, Intel, and the State of California, my work sits where AI strategy, product, architecture, and governance meet — turning human context into systems worth building, and responsible enough to keep.

15+ years Microsoft · Intel · State of California PhD, UC Davis · STS

See the work → Get in touch

What I offer

A model predicts context. Understanding it is another job

A large language model is, at bottom, a machine for predicting context: feed it enough of the world's text and it will guess what comes next, often uncannily well. What it can't do is sit in the room — read the situation, notice the hesitation, and grasp why a caseworker quietly distrusts the tool built to help her. Prediction is not understanding, and the space between them is where most of the real risk in deploying AI lives.

Closing that space is the work, and it takes both hands. Qualitative fieldwork supplies the thick, situated context a model flattens; quantitative analysis at scale shows how far a finding travels; each keeps the other honest.

Model of / model for: a model is drawn from a human context and, once deployed, reshapes it. (after Geertz)

How I think

There is no such thing as a purely technical problem

Shaheen Amir, PhD

My field is science-and-technology studies, which starts from a modest-sounding premise with large consequences: every technology is also a social artifact, carrying the assumptions, politics, and cultures of the people who make and use it. "The user" is a useful fiction, not a person. Real people, studied in their own contexts, reliably do something other than what a tool was designed for — and that gap is usually where the interesting part is.

Synthesis, in this practice, is a move from vocabulary to structure: listen for the words people reach for, cluster them, and find the smallest set of ideas that reorganizes the whole problem. From there the ambiguity is made measurable — the measures designed before the data arrives, what feels good separated from what creates value, and the evidence, rather than the loudest voice in the room, left to name the answer.

The rise of the user and the fall of people. — the title of my EPIC paper, on what industry gained, and lost, when it adopted user experience.

Where I create value

Four ways I move the work forward

Research & social science

Rigor from the human sciences

Mixed methods that actually meet — ethnography and statistics in the same study, not handed off between teams. Aimed at the hardest questions in product and research: what do people need, how would we know, and how far does the finding generalize?

15+ yrs · Microsoft · Intel · UC Davis · 4 peer-reviewed publications

Responsible AI & governance

Getting AI adopted, responsibly

Building a capable model is one problem; getting real people to trust and use it well is another. Human-in-the-loop evaluation of frontier models against actual enterprise and government workflows — the bridge between a product that works in a demo and one that's responsibly, usefully adopted at scale.

$10M+ human-in-the-loop AI pilots · 80+ use cases · 30,000+ government users

Strategy & direction

Clarity that changes the decision

Research earns its keep when it changes a decision. The craft is turning messy evidence into the one insight a leadership team can actually act on — the kind that moves budget, redraws a roadmap, and, now and then, shows up on the bottom line.

Shaped C-suite investment across $100M+ product verticals

Product & program management

Herding the cats, shipping the thing

Directing matrixed teams — and, occasionally, rival organizations — toward one roadmap; standing up the participant pipelines, panels, and incentive systems that feed evaluation at scale; keeping vendors, budgets, and timelines honest. The machinery that turns an insight into something people actually use.

50+ studies · panels & pipelines across UserTesting, Qualtrics, Ethnio · $5M+ vendor budgets

A mix of methods is only the means. The point is a story that bridges people and machine — and the foresight to turn it into lasting value.

Selected work

Four studies that drove what a team did next

One theme crosses twenty years and three industries: the future of work — how people collaborate, decide, and get things done as the tools around them change.

State of California
Office of Data & Innovation
2023 – present
Responsible AI · Human-in-the-loop evaluation

Evaluating generative AI for the public interest

California's Executive Order on generative AI arrived with a mandate and no playbook. What grew from it is a statewide, human-in-the-loop evaluation practice — testing frontier models (OpenAI, Anthropic, Gemini, Copilot, Amazon Q) against the real workflows of state analysts across dozens of departments, with staff trained to run their own studies so the capability outlasts any one mandate. AI is both the subject and the tool here: models help code transcripts and cluster first-pass themes, with a human judgment gate on every conclusion. Responsible AI as practice, not slogan — evidence about how these tools actually serve the people who depend on them.

$10M+ in human-in-the-loop AI pilots · 80+ use cases · 30,000+ state employees
validation ~50% faster · ~$200M in cost savings to date

Read the full study →

Microsoft
Microsoft 365 · Teams
2019 – 2023
Product strategy · Cross-functional leadership

One roadmap for two rival platforms

The pandemic turned collaboration software into the front line of the future of work, and Teams and Yammer overlapped in ways that confused users and set two product orgs against each other. A qual-to-quant program across twelve companies reframed the debate around a single axis — group size, or "circles of trust" — which explained which app people reached for and where the real gaps sat. The hard part was never the users; it was persuading rival leadership to trust one set of findings. The alignment that followed shipped a design that respected how people already worked.

Teams MAU +150% · Yammer/Viva +66% in 12 months
engagement KPIs up 12–20% · two rival orgs aligned on one roadmap

Read the full study →

Intel
Global Innovation Group
2011 – 2016
Foundational R&D · Executive strategy

Defining "collaboration" before the product existed

Long before "hybrid work" was a phrase, Intel's PC group had argued itself to a standstill over what "collaboration" even meant. Eight weeks took the team from ethnographic fieldwork in real offices to a 6,000-person concept test, and reframed collaboration as a human process that technology supports rather than drives — an early, foundational read on the future of work. The work anchored a new company-wide UX roadmap and helped seed a joint venture with Microsoft: the Surface Hub, which went on to become a multi-billion-dollar product line.

4 patents · seeded Microsoft Surface Hub — a later multi-$B line · 5 C-suite investments shaped
$100M+ revenue verticals · time-to-market cut 20%

Read the full study →

Microsoft & State of California
Research operations
Built twice, from zero
Program & PM leadership

Building the machinery that makes research scale

Good research runs on invisible infrastructure — and building it is as much a program-management job as a research one. At Microsoft, that meant building two custom in-house tools — for recruiting and panel management — behind an in-house research panel drawn from a mix of sources, so teams stopped rebuilding the same pipeline for every study; at the State of California, it meant creating a research-operations practice from scratch — an 8,000-participant panel with screeners, consent, and incentive systems across UserTesting, Qualtrics, and Ethnio — built to reach residents a public service can't afford to miss, and to feed AI evaluation at scale. Two organizations, two operations custom built.

2 custom in-house research tools · recruiting 50% faster, first contact to accept
an 8,000-person state research panel, built from zero

Read the full study →

A ledger of outcomes

The work, in figures

$10M+

in human-in-the-loop AI evaluation pilots led

State of California

30,000+

government employees served by the AI evaluations

State of California

280M+

people on the products the research served

Microsoft

Multi-$B

Microsoft Surface Hub — the product line this research helped seed

Intel → Microsoft

+150%

Microsoft Teams monthly active user growth in 12 months

Microsoft

~80

professionals trained in applied research methods

UC Davis & ODI

Experience

Where I've done the work

2023 — Present

Office of Data & Innovation · State of California

Program Director, Research & AI Evaluation

Governor-appointed to build the state's first human-centered AI evaluation practice: human-in-the-loop pilots across 80+ use cases and 30,000+ employees, $10M+ in AI programs, and research operations standardized across 200 departments.

2019 — 2023

Microsoft

Senior Research & Program Manager — Teams & Microsoft 365

Owned panel strategy and research operations for 50+ mixed-method studies behind products with 280M+ MAU; directed the Teams–Yammer integration and connected insight to 12–20% gains in engagement KPIs.

2018 — 2019

HumanFront.io

Founder & Principal

An ethnographic research consultancy helping organizations build people-centric innovation, product, and strategy practices.

2016 — 2018

University of California, Davis

Senior Lecturer

Taught 400+ students across eight UX-research, data-science, and science-and-technology-studies courses (92% satisfaction); mentored 30+ into top technology firms.

2011 — 2016

Intel Corporation

Lead UX Researcher — Global / Future Innovation Group

Drove foundational research across $100M+ revenue verticals, shaping five C-suite investments (big-data AI, AR/VR, HPC) and cutting time-to-market 20% across the innovation pipeline.

Writing & speaking

On innovation and the social life of technology

I publish, teach, and speak internationally on innovation, ethnography, and responsible AI. Scholarly work is published as Shaheen Amirebrahimi.

2026
The Moral Architecture of Innovation
Essay · on ethics and governance in public-sector innovation
2018
Innovate or Die: Silicon Valley and the Tribes that Built the Future
Doctoral ethnography · four years of fieldwork on the rise of UX · Library of Congress / ProQuest
2017
Time for a Digital Detox: Burnout, Addiction, and Desperation in Silicon Valley
EPIC Proceedings
2016
The Rise of the User and the Fall of People: Ethnographic Cooptation and a New Language of Globalization
EPIC Proceedings
2015
Moments of Disjuncture: The Value of Corporate Ethnography in the Research Industrial Complex
EPIC Proceedings

Boards

Conference ChairEPIC 2026 · Chicago

Paper Committee ChairEPIC · Los Angeles 2024

Capabilities

Methods, craft & leadership

Ethnographic & qualitative

In-depth interviews
Contextual inquiry & field studies
Ethnographic research
Diary studies
Moderated usability testing
Card sorts & participatory design
Heuristic evaluation

Quantitative & computational

Surveys (SUS, CSAT, NPS)
Usability benchmarking
A/B testing
Behavioral & clickstream analytics
Statistical analysis (t-tests, regression)
AI & model evaluation (human-in-the-loop)
Conjoint & value modeling

Product, program & org leadership

Product & program management
Research operations & participant pipelines
Directing matrixed & cross-functional teams
Research strategy & democratization
Vendor & budget management ($5M+)
Executive alignment & communication
Coaching & team building

Domains

The future of work · Enterprise collaboration · Public-sector services · Responsible AI · Emerging tech (AI, cloud, AR/VR)

Tools

UserTesting · User Interviews · dScout · Dovetail · Qualtrics · Ethnio · Airtable · R · SPSS · SQL

Contact

Let's put people at the center of what you're building.

Contact me → LinkedIn Publications (EPIC) Download CV

Shaheen Amir · Based in California PhD, Cultural Anthropology / STS · UC Davis