Machine Learning Resident - Client: Fillip Fleet; term

Machine Learning Resident - Client:
Fillip Fleet (12 month term)

“If you are excited about applying to LLMs to tackle real-world challenges in fleet operations and financial spend management , this is a perfect opportunity for you. Be a part of the team of research and machine learning scientists building deployable real-world ML applications from ground up and get mentored by some of the best minds in AI during the process.”

- Kunwar Saaim, Machine Learning Scientist and Amor Provins, Product Owner, Advanced Technology

This is a paid Residency that will be undertaken over a twelve-month period with the potential to be hired by our client, Fillip Fleet, afterwards (note: at the discretion of the client). The Resident will report to an Amii Scientist and regularly consult with the client team to share insights and engage in knowledge transfer activities.

Successful candidates will be members of a cross-functional project team with backgrounds in ML research, project management, software engineering, and new product development. This is a rare opportunity to be mentored by world-class scientists and to develop something truly impactful.

Fillip Fleet serves as the modern intelligence layer in a traditionally complex and dashboard-centric fleet card management market where operators face significant manual overhead. The primary commercial driver is to streamline this process by turning natural language business intent into immediate database updates, helping fleet managers save time and minimize operational errors.

A central objective is to eliminate the cognitive burden of navigating complex dashboard menus, specifically targeting the manual bottleneck where operators must individually update each fleet card policy. This automation is designed to enable fleet managers to execute bulk changes seamlessly (e.g., "restrict fuel purchases to weekday working hours for the Vancouver team") while maintaining high standards for precision, data isolation, and operational efficiency.

Amii collaborated with Fillip Fleet to analyze the foundational business challenge—simplifying fleet card policy creation and updates—and investigated scalable ways to approach the problem using Agentic AI workflows. The review focused on the technical integration of Large Language Models (LLMs) with Fillip Fleet's database architecture to ensure secure, efficient, and reliable operations. This comprehensive analysis addresses the limitations of the current system, introduces an advanced architecture for dynamic rule-based configuration, and outlines a secure implementation path using the Model Context Protocol (MCP)

The project will deliver an intelligent system that shifts the operational burden away from manual menu navigation. Instead, fleet managers will be able to dictate business intent using natural language (e.g., "Restrict fuel purchases to weekday working hours for the Vancouver team"), which the system will automatically interpret, verify, and safely execute across multiple policies simultaneously.

The implementation follows a structured, two-phased approach:

• Phase 1 (The Existing System):
  Developing the natural language interface and specialized sub‑agents (Routing, Query, and Update agents) to manage bulk changes across the current structure of fleet defaults and individual overrides.
• Phase 2 (The Intermediate Layer Architecture):
  Transitioning the backend database logic to a rule‑based, attribute‑driven segment layer. This allows policies to apply dynamically to any users matching specific conditions (such as location or vehicle type), permanently eliminating duplicate manual exceptions and ensuring scalable rule execution.

Are you passionate about building great solutions? You’ll be presented with opportunities to both personally and professionally develop as you build your career. We’re looking for a talented and enthusiastic individual with a solid background in machine learning, large language models, and agentic AI systems, along with proven experience in applied settings.

• Design, build, and evaluate a multi‑agent natural…