AgentFit — Workflow Opportunity Mapper
An interactive product that scores where an AI agent belongs, how autonomous it should be, and what system/control pattern fits the workflow.
Production AI, 0→1 financial infrastructure, and measurable operating leverage — built for workflows where speed matters and trust matters more.
Production agents, AI product judgment, and financial infrastructure. Start with the flagship stories; the deeper systems track record sits underneath.
An interactive product that scores where an AI agent belongs, how autonomous it should be, and what system/control pattern fits the workflow.
Move the inputs. AgentFit turns workflow characteristics into an illustrative fit score, autonomy recommendation, control pattern, and capacity estimate.
Four additional 0→1 / production systems across optimization, validation, FX, and margin.
Employer work is intentionally generalized. Internal implementation details, client information, and sensitive system specifics are omitted; visible product names are contextual only.
My default is not “put an LLM in the loop.” It is: understand the workflow, expose the right tools, bound the action space, evaluate behavior, and prove economic value.
Reusable MCP servers, domain APIs, Pydantic-typed actions, deterministic services, human approval, and evaluation / QA baselines.
Map the hidden handoffs, constraints, exceptions, and failure costs before choosing the model.
Separate model judgment from deterministic logic and make consequential actions observable and reviewable.
Track whether the product returns capacity, improves adoption, changes unit economics, or creates revenue.
I write short notes on AI product economics, agent controls, fintech infrastructure, and the product decisions underneath new technology.
The best AI product may not use the best model everywhere. Routing intelligence is increasingly a product and unit-economics decision.
The useful question is not whether an agent can act. It is what evidence, reversibility, and controls justify letting it act.
In high-stakes finance, the model is only one layer. Data quality, entitlements, APIs, controls, and workflow design determine whether it ships.
I build AI agents and financial infrastructure for complex, high-stakes workflows.
Lead PM at Quantile (LSEG), shipping production AI agents and 0→1 financial products. My work has cut expert workflows from hours to minutes, scaled operations without added headcount, and generated $3M+ in new and expansion ARR.
Previously built pre-trade margin and derivatives analytics at OpenGamma.
PROFILE / 2026B.A. Economics & Linguistics · 3.9/4.0
Cum Laude · Linguistics High Honors
Acquired by London Stock Exchange Group
Shipping production AI agents and reusable human-supervised architecture across enterprise financial workflows.
Owned discovery, strategy, launch, and GTM across rates, FX, and cross-currency optimization; generated $3M+ in new and expansion ARR.
Acquired by Trading Technologies
Originated and shipped a 0→1 pre-trade margin simulator and led discovery across trading, treasury, risk, and operations, shaping capabilities across 20+ enterprise clients.
PRODUCT · AI AGENTS · FINTECH · INFRASTRUCTURE
If you’re building AI-native products, financial infrastructure, or software that turns complicated enterprise workflows into something dramatically better, I’d love to talk.