AI Engineer · Melbourne

Liam
Harris

liamharris1199@gmail.com+61 437 796 370linkedin.com/in/liamcharris

Skills

Python · Django · Elasticsearch · AWS · PostgreSQL · Redis · Docker · OpenAI & Anthropic SDKs · MCP · RAG · LLM evaluation · prompt optimisation · agent orchestration

Experience

Affinda

Melbourne

Melbourne startup building intelligent document processing products powered by LLMs.

AI Engineer

Oct 2024 – present
  • Led the design and build of the Affinda Agent — a configuration and onboarding agent with access to every action a user can take on the platform. Built it end-to-end: MCP server, agent gateway, prompts, tool use, retrieval, memory, and orchestration. Drove the rebuild of the application and onboarding experience around it, and owned all evals and experiments behind its production behaviour.
  • Established Affinda's LLM evaluation and feedback infrastructure from the ground up. Evals have directly driven major architectural decisions — including a full teardown and rebuild of the document splitting approach.
  • Own prompt optimisation across core pipelines — evolutionary algorithms for prompt evolution and continuous experimentation across prompting, model routing, and orchestration that feeds back into product decisions.
  • Drive internal AI tooling and adoption, including custom harnesses built on top of coding agents to work around limitations introduced by their RL post-training.
  • Partner closely with product, frontend, and operations to own the full idea → implementation → measurement → iteration loop on AI features in production.

Junior AI Engineer

Jan 2023 – Sep 2024
  • Designed Affinda's retrieval fingerprinting algorithm — the core embedding mechanism behind the platform's instant-learning customer experience.
  • Built bespoke text extraction and classification systems for high-volume document processing in medical research, insurance, and accounting.

AI Solutions Analyst

Jun 2022 – Dec 2022
  • Drove rapid prototyping of a new document processing product, owning scope, competitor analysis, and model evaluation alongside the AI team.

Education

University of Melbourne

Dec 2022

Master of Engineering (Mechatronics)

First Class Honours · WAM 85.4 / 100. Coursework in AI, Computer Vision, Robotics, and Autonomous Navigation.

University of Pennsylvania

Fall 2019

Exchange — Wharton and School of Engineering and Applied Science

University of Melbourne

Dec 2020

Bachelor of Science (Mechatronics Systems)

First Class Honours · WAM 88.5 / 100

Projects

Rumi

[In development]

Mobile app for finding flatmates and managing sharehouses, targeting the Melbourne rental market. Co-founded and leading engineering: native iOS app in Swift, Django backend with Django Channels and Redis powering real-time group chat, Hinge-style browsing with personality prompts, and AI-powered applicant matching against sharehouse profiles. Also contributing to product strategy, marketing, and business.

UniMelb Rover Team — Autonomous Navigation (Capstone)

Designed a high-performance localisation system using ultra-wideband (UWB) sensors, optimised sensor placement algorithms, and AI localisation models in Python, TensorFlow, and MATLAB.

Honours & Awards

  • Dean's Honours List
  • Boeing Award for Excellence
  • Melbourne Chancellor's Scholarship
  • Melbourne Global Scholars Award
  • Ormond College Scholar
  • Startmate Student Fellow
  • Dux of School — All Saints Anglican School, 99.95 ATAR