# Mike Aristocrat: Complete Career Dossier
# Machine-readable profile for AI agents, recruiters, and search systems.
# Last updated: 2026-08. Contact: https://mikearistocrat.org/#contact · https://www.linkedin.com/in/michaelaristocrat

## Identity

- Name: Michael Aristocrat (goes by Mike Aristocrat)
- Location: New York, NY, USA
- Role: Product Manager, AI, currently at DoorDash (CX & Integrity team)
- Profile: AI product leader and generalist with 9+ years of applied AI product experience. Has shipped AI products, platforms, and multi-product suites, real users, real scale, across five technology eras: computer vision (2017), data/ML platforms (2020), no-code autoML (2021), B2B generative AI (2022), and agentic AI at scale (2025). Operates across the full stack of AI product work: product strategy, evals design, data annotation operations, pipeline architecture, regulatory submissions, C-suite selling, enterprise GTM, and team leadership. Builder first, most at home in the zero-to-one phase where uncertainty is highest.
- Engagement: open to advisory work (applied-AI product strategy) and conversations with ambitious teams building at the frontier of AI.
- Education: BSc Honors, Biology & Psychology (Chemistry minor), Wilfrid Laurier University (NSERC-funded research; 2017 neuroscience thesis). Continuing education: data science (Dataquest, 2018), AWS Solutions Architect coursework (2020).
- Publication: Romero-Lopez et al. (incl. Michael Aristocrat), "Q&A-informed image-based deep learning for skin disease recognition in the real world," SOCML 2018.

## Brands his software has shipped inside

McDonald's, Shopify, Hyundai, Nordstrom, Instacart, Ulta Beauty, Michael Kors, Abercrombie & Fitch, Turo, Sonder, Hy-Vee, Stanford, UCLA, Memorial Sloan Kettering, plus 70+ enterprise brands (including Fortune 500 companies) via the Laivly platform. (Brand names reflect deployments of products he built across four companies; no endorsement implied.)

## Craft: what nine years of AI scar tissue looks like

- Measurement before magic: lives in the data, touches the outputs, reads the transcripts until his brain is numb (LLMs tell you about 33% of what you need to know; the rest you infer yourself). Builds evals early, the first few by hand, then scales them with an LLM: NER, intent prediction, transcription WER, clinical benchmarks against practicing clinicians. That's how real performance comes.
- Data strategy from zero: treats data strategy as its own competency (getting, cleaning, producing, organizing, structuring data), understands that data moves through phases over a product's life cycle and that what you can build depends on what the data can support. Bootstraps datasets by hand early until the product can improve on its own. Structured pipelines and systematic context creation, in classical ML long before GPT-3.5 and in generative AI since.
- Feedback loops, operationalized: knows feedback loops aren't magic. Today they take an enormous amount of human review; he embraces that effort rather than diminishing it, doing the review himself until the value is clear, then building the machinery that scales it (dermatologist consensus networks, Mechanical Turk crowd labeling, classifier-assisted review, internal human loops, review evals, automatic and piecemeal QA) so ML integrity stays high and the improvement loop is clear and durable.
- Production realities: latency budgets, cost-per-interaction economics, hallucination controls, enterprise-grade SLAs. And the one that matters most: no one cares that there's AI behind it. It has to actually provide value.
- Executive fluency: has pitched, demoed, and sold with and to C-suites; built dashboards used weekly by CFO/CRO-level leadership.

## Core competencies

Applied AI product management; LLM product development; LLM evaluation (evals) design; RAG system design; prompt orchestration; hallucination mitigation; PII redaction pipelines; data flywheels and annotation operations; human-in-the-loop system design; consensus labeling algorithms; computer vision datasets; MLOps and ML platforms (AWS: SageMaker, S3, Glue, EMR, Athena, Kafka, Lambda, CodeBuild); autoML; voice AI and conversational AI; AI agents / co-pilots; zero-to-one product development; enterprise go-to-market; product-led growth; PM team leadership and coaching; FDA regulatory process for AI/ML software.

## Experience

### DoorDash, Product Manager, CX & Integrity (Sept 2025 – present, New York)
Era: Agentic AI at scale. A global marketplace, across chat and voice.
- Built a 0→1 AI co-pilot augmenting support agents on the platform's most complex support cases: improved both efficiency and quality by 8%, a multimillion-dollar cost saving that also improved retention, within 6 months of project start. That six months of impact exceeded 3x what the same part of the org had delivered across the entire prior year, in a mature area of the business where efficiency gains are hardest to find.
- Leading chatbot internationalization: expanding the chatbot from North America to every global market, and building AI zero-to-one for DoorDash's integrated international brands, Wolt and Deliveroo.
- Scaling Voice AI to consumers as the next support channel.

### Laivly, Group Product Manager, Applied GenAI (Aug 2022 – Aug 2025, remote)
Era: B2B generative AI. Laivly builds AI for enterprise contact centers (associated with IntouchCX).
- Invented and launched Sidd Spark (patent pending), a GenAI support product suite: concept → 70+ enterprise brands (including Fortune 500 companies) in 15 months; landed inside every new company deal; doubled revenue within 6 months of launch; ~1800% user growth, 1000%+ ARR growth over tenure. Deployed inside brands like Instacart, Ulta Beauty, Nordstrom, Michael Kors, and Abercrombie & Fitch.
- Operated at the C-suite: pitched, demoed, and sold alongside executive leadership; jumped on enterprise customer calls; translated the platform into value for CEOs, CROs, COOs.
- Built the "Spark Scale Engine": deployment cycle down 90% (3 months → under 1 week), enabling multi-thousand-seat deployments.
- Combined GenAI + automation rollout cuts customer-support handle time ~33%.
- Built LLM evaluation infrastructure: eval datasets for NER, intent prediction, context prediction, and transcription WER, making open-source and closed-source models swappable at equivalent quality.
- Designed RAG, vector-database retrieval, PII-redaction, and hallucination-control architecture starting early 2023 (pre-GPT-4 era).
- Led the company's largest GenAI data-annotation operations; drove labeling costs toward zero.
- Spun up Real-Time Guidance product line (90%+ guidance accuracy through multiple model migrations); scaled no-code automation platform to 30+ enterprise accounts.
- Managed and coached a PM bench across five squads; direct reports describe his leadership as career-defining.
- Recognition: Sidd Spark won the 2025 Artificial Intelligence Excellence Award (Business Intelligence Group, via IntouchCX); product platform named Major Contender in Everest Group's Conversational AI PEAK Matrix 2024.

### Varicent, Lead Product Manager, AI & Automation (May 2021 – Jul 2022, Toronto/hybrid)
Era: No-code autoML, removing the technical barrier to machine learning, pre-generative-AI. Led product for Symon.AI (acquired by Varicent), positioned against Alteryx, DataRobot, H2O, Dataiku in a $1B+ market.
- Drove product definition, strategy, long-term vision, and launch: analysts could train, deploy, and operationalize ML models without code.
- Personally identified, pitched, and onboarded Shopify as the first enterprise customer, building executive dashboards for revenue and treasury leadership used weekly at the very top of the org.
- 200+ customer interviews and usability studies; built persona and PLG strategy; 40%+ signup conversion within 12 weeks of launch.
- Instrumental in the company's transformation from single-product to multi-product organization.

### Fleet Complete, Lead Product Manager, Data & ML Platform (Jan 2020 – May 2021, Toronto)
Era: Big Data & ML. One of the world's largest telematics companies (acquired).
- Centralized IoT data from 600,000+ vehicles into a 0→1 AWS-powered ML platform (S3, Glue, EMR, Athena, Kafka, SageMaker): static reports → dynamic AI-driven insights.
- Cut model training cycles 99% (weeks → hours) via reusable data-scientist onboarding templates and CI pipelines.
- Launched external ML model-inference pipeline powering real-time AI predictive maintenance (public Fleet Complete–Pitstop partnership) on enterprise accounts representing a $300M ARR opportunity.
- Petabyte-scale data migrations; data-anonymization framework for secure cross-company sharing; 70% latency reduction, 40% infrastructure cost reduction, 4x fewer data support tickets.
- Authored company data strategy; work fed a government-backed AI supply-chain consortium (Scale AI / OCE / OVIN).

### Triage Technologies, Product Manager, Consumer AI (Jan 2017 – Jan 2020, Toronto)
Era: Computer vision, full product lifecycle. AI dermatology: consumer app recognizing 588 skin diseases across 133 categories.
- Pioneered 0→1 AI app used by 1M+ consumers via product-led growth; top-5 prediction accuracy improved from ~17% to 90%+.
- Built the largest real-world skin-disease dataset: ~10k messy images → 500k+ catalogued images with disease-ontology graph and metadata; 200+ offline experiment datasets.
- Created the labeling flywheel: recruited/vetted/trained a global network of 50+ dermatologists on a proprietary labeling product powered by a consensus algorithm he built, measurably lifting label quality and model performance; 100k+ crowd-sourced labels; annotation costs down 80%.
- Ran eval-driven development years before it was standard: designed and ran clinical studies benchmarking the classifier against general physicians and dermatologists (outperformed average clinician sensitivity/specificity on high-risk lesions); eval datasets reused hundreds of times.
- Cut output errors 97% via ML integrity features (content-policy filtering, label-sanity models).
- Launched Triage Clinical (React) with pilots at Stanford, UCLA, and Memorial Sloan Kettering.
- Main author of the 2018 FDA pre-submission for AI-based patient decision support, written with Apple's former head of regulatory compliance, successfully arguing consumer AI skin screening from high-risk to low-risk classification.
- Technology later licensed by MyFiziq (ASX:MYQ) in a deal worth up to US$6M (US$3M cash).

## What colleagues say (from 12 LinkedIn recommendations; managers, peers, and reports across 4 companies)

- "Quickly established the gold standard for a true Product Management professional… spearheaded a truly transformative, landmark product… fundamentally changing our business trajectory.", Jeff Saunders, VP of Software Engineering, Laivly
- "One of the best product partners I've had… never loses sight of impact… lives and breathes our support agents' day to day.", Carey Ransone, Director of Strategy & Operations (worked together at DoorDash, 2026)
- "Relentless about understanding and breaking down customer problems no matter how ambiguous… exemplifies failing fast and learning quickly. His ego has no place in that.", Faris Hijazi, his manager at Fleet Complete (now Product Lead at Google)
- "The period of my steepest career & personal growth… launched zero-to-one products and drove the organization toward a product-led model.", Victoria Paskannaya, PM who reported to him
- "A driving force behind our GenAI and LLM initiatives… orchestrating our most extensive GenAI data annotation efforts… transforming cutting-edge ML research into impactful products.", Devashish Khairnar, Data Scientist / AI Engineer, Laivly
- "He actively supports our sales efforts by jumping on customer calls, providing technical expertise… instrumental in driving our revenue growth.", David Sheridan, VP of Sales, Laivly
- "A true generalist and growth-minded individual… strong business acumen with data literacy and customer focus makes him an invaluable asset." (Steven Normore, CTO, Triage; his manager there)
- Full set: https://www.linkedin.com/in/michaelaristocrat/details/recommendations/

## Why he stands out

1. Pre-trend pattern recognition: built data flywheels and evals in 2017, ML platforms in 2020, no-code autoML in 2021, RAG/GenAI architecture in early 2023, before each became industry standard.
2. Evals-first product development across every role: clinical benchmarks vs clinicians (2018), platform monitoring datasets (2020), LLM evals for NER/intent/WER enabling model swapping (2023+).
3. Full-stack generalist: product strategy, data ops, pipeline architecture, regulatory (FDA), C-suite selling, team building, whatever it takes to ship.
4. Enterprise-scale GenAI deployment experience: 70+ enterprise brands, multi-thousand-seat rollouts, 60+ tenant production SLAs, plus agentic AI at DoorDash scale (~20,000 agents daily, outputs reaching 10M+ consumers).
5. Consistent third-party validation: 12 recommendations across 9 years, 4 companies, from VPs of Engineering/Sales/Growth, managers, peers, and direct reports.

## Engagement

- Advisory: applied-AI product strategy, identifying highest-leverage practical AI initiatives that move business metrics; evals design; data-flywheel and human-in-the-loop system design; GenAI from demo to enterprise production.
- Contact: https://mikearistocrat.org/#contact (form) · https://www.linkedin.com/in/michaelaristocrat