The practice
Twelve senior data scientists, AI engineers, and researchers.
These are the people who make up the practice. Each profile below is one person, described by the work they personally designed and built over their careers. Tell Shaun the problem. Shaun names who on this page has solved that shape of problem before, and you meet them to scope the work. Before any access to your systems, the named individuals are provided to your security team.
Partners
Led by its partners
Shaun Applegate-Swanson, Co-Founder & Partner
Mark Cicoria, Co-Founder & Partner
Jason Larkin, Partner
The first conversation is thirty minutes with Shaun, on one problem you name. Shaun matches the problem to our consultants and stands behind the engagement.
The practice
Between them they cover ten sectors.
C-01
Shaun Applegate-Swanson, Co-Founder & Partner
Early data scientist three times, each through acquisition
- First or second data scientist hired at three VC-backed startups (~$150M combined funding), all three acquired: Radius → Kabbage/Amex, CloudPhysics → HPE, CloudKnox → Microsoft
- Stayed on post-acquisition at Microsoft as Principal Data Scientist
- Five patents in least-privilege security and permissions risk: two issued, three pending
- Built Scala/Spark pipelines converting 100 billion data points a day into modeling assets
- BS Physics + BS Mechanical Engineering (Carnegie Mellon); MS Mechanical Engineering (Caltech)
C-02
Mark Cicoria, Co-Founder & Partner
GPU cloud observability, built end to end; serial founder and 500 Startups alum
- Designed and built the observability platform for a GPU cloud provider, unifying scheduler, GPU, and node telemetry, out-of-band management, network fabric, storage, tracing, and logs into one multi-tenant system
- Built the diagnostics interface used by that provider's AI researchers
- Co-founded or led technology at seven companies: took one through its acquisition, built the technology another sold, and grew a team-collaboration SaaS to 200+ customers
- Built low-latency blockchain trading systems running $5M+ in annual volume
- Five years at the U.S. Department of Energy building auditable, compliance-grade systems
C-03
Jason Larkin, Partner
Quantum computing researcher: machine learning for oncology, and PHI under post-quantum security
- Architected quantum-classical hybrid machine learning for oncology diagnostics
- Built a Protected Health Information platform integrating post-quantum cryptography, quantum key distribution, and homomorphic encryption
- Advised a CEO on quantum strategy for drug discovery, and ran technical due diligence on early-stage quantum companies for VC investment decisions
- Seven years as Senior Research Scientist at Carnegie Mellon's Software Engineering Institute
- PhD Mechanical Engineering, Carnegie Mellon; published across thermal transport, turbulence, and formally verified robot and vehicle control
C-04
Agentic AI for financial crime, labeled prototype and trial
- Built an enterprise AI vendor's first LLM-based multi-agent anti-money-laundering prototype, used in Fortune 100 sales demos and roadmap planning
- Built an agentic system automating a Fortune 100 card issuer's fraud and sanctions investigation workflows, run as a trial rather than taken through to production
- At C3.ai: predictive maintenance for US Air Force E-3 Sentry aircraft, and money-laundering detection for a major international bank
- Designed the curriculum that trained and onboarded 15 new data scientists
- Began at Morgan Stanley, managing execution of a US$376MM Hong Kong IPO
- MS Computational Science & Engineering (Harvard); BSc Finance (Wharton)
C-05
Moves engineering teams to shipping with agents daily
- 25+ years across enterprise, agencies, and startups they founded
- Led 14+ engineers across four product workstreams from AI-curious to shipping with agents daily
- Held a hard standard: agents work test-first, everything verified with tests, linting, and guardrails, then code-reviewed, often by a different model than the one that wrote it
- Shipped two iOS apps to the App Store built almost entirely with agents, in roughly two and four weeks
- AWS Certified Solutions Architect and AI Practitioner
C-06
Security built into finance-grade systems
- Patented inventor in cryptographic authentication of physical assets
- Creator and maintainer of a widely adopted open-source MongoDB schema analyzer (~1.8k GitHub stars, used at Microsoft's FUSE Labs)
- Runs finance-grade Go and Rust systems on multi-region AWS
- Security-by-architecture: highest-sensitivity secrets on a no-IP host, egress via dedicated proxy, systemd sandboxing, and open-source metrics and log-aggregation telemetry
- CTO or founding engineer at four companies; created and taught a graduate course at NYU's ITP
C-07
How do we know the system works? Evaluation at platform scale
- Works the harder question: not “what is the data saying” but “how do we know our systems are actually working”
- At Meta, built the quality measurement infrastructure (accuracy, precision, recall) for a global content review operation
- Seven-plus years at Twitter and X designing graph-based coordinated-behavior detection
- Contributed rule logic to an enforcement engine processing billions of daily events across hundreds of millions of monthly active users
- Prior career in finance and risk starting at Goldman Sachs
C-08
Teaches generative AI to working teams
- Co-founded one of the first full-time data science immersive programs
- VP of Academic Excellence at Galvanize
- Eleven years authoring for Pearson on generative AI, deep learning, and distributed computing, including a title on generative AI with large multimodal models in PyTorch and Hugging Face
- Ran data science training for a Fortune 100 company
- Taught at Strata, PyData, and DataWeek
C-09
Large rewrites, cut over with zero reported bugs
- At PayPal, led the rewrite of the server-driven UI framework powering the majority of the app's key screens
- Replaced the model layer with compiler-enforced type-safe Swift
- Built a differential test harness proving output equivalence between old and new, enabling a single feature-flag cutover to production with zero bugs reported
- A decade of consumer iOS product depth, now pointed at AI-native applications
- Built the mobile stock trading framework behind Charles Schwab's iPhone app
C-10
Analytics teams built inside health systems
- Established the Predictive Analytics, ML, and App Development team at Texas Children's Hospital
- Certified on the dominant electronic health record platform, implementing at multi-hospital, academic, and for-profit health systems
- Built a unified Analytics department (BI, data science, data engineering, and research) at findhelp, rising to Director
- Specialist in standing up analytics organizations inside large regulated environments
C-11
Grid forecasting and DER modeling in production
Unavailable Not taking new work right now. Check back in December.
- Grew a grid-data startup's data science team to 12
- Lead developer on the CPUC High DER Electrification Impact Study, coordinating many simultaneous workstreams
- Built and ran a bottom-up grid forecasting pipeline
- Deepest grid expertise on the team: demand and generation forecasting, AMI and SCADA, DER adoption modeling, grid graph modeling
- Published research on accelerating rooftop solar permit approvals
- Previously led the Graph team at Radius and ran every production ML model at Kabbage
C-12
Scientific rigor from evolutionary biology and complexity science
Unavailable Not taking new work right now. Check back in December.
- Stanford PhD in Ecology and Evolutionary Biology
- NSF Postdoctoral Research Fellow at UC Davis, researching the evolution of gene expression in ants and bees
- Founded or organized three research communities of practice
- Currently teaches biology as associate faculty at a community college
- Works in active inference, cognitive security, and systems thinking, suited to a client who wants scientific rigor and framing rather than production engineering
Each profile is one person's own work, in the roles and at the employers where they did it. Their achievement, not the firm's. Years count technology, engineering, data science, and research roles only; patents are the individuals', not Transmute Data's. We hold no HIPAA attestation of our own — we work inside your governed environment, under your BAA.
How you meet them
Meet the person before the work starts
Tell Shaun the problem and the shape of the team. Shaun names who fits, and you meet them before the work starts.
A principal leads, and stays
Every engagement is led by a principal-level data scientist who stays through delivery. Specialists from the twelve join as the scope needs them.
A 30-minute call, on one problem you name.
Tell Shaun the problem on the call. Shaun names who fits, and you meet them before any work starts.