AI Developer

Marsh McLennan•Toronto, ON

3d•Hybrid

About The Position

We are seeking an AI Developer to design, build, and maintain scalable backend services and APIs in Python, with a strong focus on Flask and FastAPI, including both synchronous and asynchronous patterns. The position also involves developing production-grade AI and LLM capabilities such as REST and WebSocket APIs, RAG pipelines, agent/tool-calling workflows, and model evaluation and optimization across GPT and open-source solutions.

Requirements

3+ years of experience in a corporate IT environment
1+ year of experience in Artificial Intelligence, Machine Learning, Python, Flask, and FastAPI; sound software design, debugging, and performance profiling skills.
Solid SQL Server & PostgreSQL knowledge (schema design, indexing, optimization; ORM experience).
Maintaining adherence to security, privacy, and compliance best practices (secret management, data handling, access control).
Strong commitment to secure, reliable, well-documented code ensuring adherence with quality standards.

Nice To Haves

Exposure to containerization and orchestration (Docker, Kubernetes) for deployment workflows.
CI/CD familiarity (e.g., GitHub Actions) for automated testing, build, and deployment pipelines.
Light JavaScript for integration or tooling.
Organized, self-directed, and deadline-focused in distributed team environments.
Fast learner, quality conscious and committed to deadlines

Responsibilities

Designing, building, and maintaining robust, scalable backend services and APIs in Python (focus on Flask and FastAPI, synchronous and asynchronous patterns).
Writing clean, efficient, and scalable Python code with strong attention to performance, reliability, and security.
Developing and exposing AI/LLM capabilities as production-grade APIs (REST / WebSocket streaming responses, authentication).
Implementing Large Language Model (LLM) solutions (GPT family and open-source models) including prompt design, evaluation, versioning, and cost/performance optimization.
Build and maintain RAG pipelines (ingestion, embeddings, vector search e.g., pgvector) and agent/tool-calling (MCP / function calling) workflows.
Integrating external/internal tools and data sources for LLM-based agents (databases, APIs, file systems, messaging platforms).
Integrate data sources (SQL Server, PostgreSQL) and implement efficient schemas, queries, and transaction handling. ORM usage e.g., SQLAlchemy).
Applying classical machine learning (scikit-learn) and statistical techniques when they provide simpler or more cost‑effective solutions
Performing systematic evaluation and monitoring of AI/LLM outputs (quality, hallucination mitigation, latency, cost KPIs) and implementing guardrails/content filtering.