Who we are.
Behind the practice is Camille Lambert: twenty-five years building software, the last eight running the people who build it, and the past year building with generative AI full time.
AI tools, products, libraries, governance
The past year has been building with generative AI full time. Multi-agent orchestration systems on AWS Bedrock, Step Functions and Lambda, with Claude as the reasoning engine. Current agent tooling: LangGraph, LangChain, RAG pipelines and evals. One AI product shipped end to end, a wellness application built in React Native and Expo.
Three AWS AI/ML certifications: Generative AI Developer (Professional), Machine Learning Engineer (Associate), and AI Practitioner. This is not new territory. Work at Citibank included ML-driven insights and a product recommendation engine, and a master's at CNAM included artificial intelligence coursework.
On AI governance, the approach comes from the data side: governance models with audit trails, access controls and compliance monitoring, and GDPR, CCPA, SOX and SOC 2 controls mapped onto data platforms. AI governance is the same discipline pointed at a newer object. We know which questions the control functions will ask because we have been the person answering them.
Software, cloud, integrations
Production code from 2001 to 2009 in C#, SQL, VBA and PL/SQL, first for government agencies, then for hedge fund analysts and risk managers at Swiss Capital Group. Later, web applications in React, Node and MongoDB, containerized and run on AWS. The Bail Project's case management system was built on Google Cloud.
Excel sits here because the Excel work has always been engineering: VBA frameworks, Excel and MatLab APIs, generated functions and reports. Not spreadsheet cleanup.
The recurring shape of the career is joining systems that do not talk to each other. At Citibank that meant trade capture, pricing, market data, Bloomberg and Intex. At The Bail Project it meant Salesforce, Google Cloud and finance.
Data functions at both ends of the size range
At Citi, the pricing, risk and P&L data behind an $85 billion book across thirteen desks. At Disney, subscriber, partnership and content data covering more than 60 million subscribers in 160 countries. At Swiss Capital and The Bail Project, the same work at boutique scale, where there is no data team and the person who builds it is the person who runs it.
Working at both ends makes the current moment obvious. Generative AI, used well, supports every step of this work: reconciliation, integrations, pipelines, observability. It collapses the cost of implementation and maintenance to a fraction of the headcount.
Product, project and program management
Product management has been the job title since 2011: seven years as a technical product manager and vice president at Citibank, then four years at Disney. The whole cycle, for traders, finance teams, analysts and executives: discovery, roadmaps, PRDs, backlogs, sprint planning, launches, production support, deprecations.
Project and program management is the same job: scope, tickets, risks and multi-team coordination, backed by a PMP, a CSM and a CSPO.
The clearest proof is The Bail Project. Hired at the organization's inception as interim CTO and CFO, Camille built two departments, a budget, and a working case management system for 3,000 clients across six states in seven months.
What happens after it is built
Stakeholder management has been most of the job. Traders, desk heads, sales, financial control, operations, risk and audit at Citibank. Engineering, analytics, data science, finance, FP&A, controllership, governance and compliance at Disney. A CEO and a board at The Bail Project.
Change management is treated as work, not a slide: usage and discrepancy analysis to find how people actually use a system, then the training, wikis and onboarding that make a change hold.
On GRC and IT risk: SOX-style controls at The Bail Project, regulatory reporting under an OCC deadline at Citibank, and data governance with audit trails and access controls, working from COBIT, ITIL and ISO 27001. We are not compliance practitioners and we do not sell certification.