Skip to content
View Philori22's full-sized avatar
🏠
Eating fruit
🏠
Eating fruit
  • University of Bath
  • Bath

Block or report Philori22

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Philori22/README.md

Philip Lorimer

Robotics ML Researcher Β· Applied AI Engineer Β· Builder
Teaching robots to see, move, and act with human-aligned intent β€” and shipping the software around them.


      β•±β•²      perception       β•±β•²β•±β•²       planning      β•±β•²β•±β•²β•±β•²      control
     β•±  β•²    ─────────────    β•±      β•²   ─────────────  β•±        β•²  ─────────
    β•±    β•²  scene Β· semanticsβ•±  goals β•²  trajectories  β•±  actions  β•²  motors
   ╱──────╲                 ╱──────────╲              ╱──────────────╲
                                                                          β†’ robot


🧠 About

I'm a Doctor of Engineering in robotics and AI (University of Bath) with peer-reviewed research in learning-based control for real-world robots. Six years of doctoral and postdoctoral research β€” including industry R&D embeds β€” followed by a move into full-time industry as a robotics and computer vision engineer.

On the side I run Eight Sprouts, a small studio where I ship AI-native products and take on applied-AI consulting work. I'm most useful when a project crosses disciplines β€” where research rigour meets real-world engineering.

πŸ”¬ Researcher

Reinforcement and imitation learning for real robots.
Sim-to-real. Evaluation rigour.
Published at IEEE/RSJ IROS.

πŸ›  Engineer

Python, C++, ROS 2, FastAPI,
Postgres, Neo4j, Azure.
Pipelines, services, hardware.

πŸš€ Builder

Agentic systems, knowledge graphs,
conversational reasoning β€”
production-grade, on the Claude API.


πŸ“š Published Work

Year Venue Paper
2024 IEEE/RSJ IROS Β· Abu Dhabi Reinforcement Learning of Dolly-in Filming Using a Ground-Based Robot
2025 IEEE RA-L Β· under review Learning Dolly-in Filming from Demonstration Using a Ground-Based Robot

🚧 What I'm building right now

Being upfront: this is a moment of transition. My postdoc wrapped in March, my industry role's remit shifted, and I've redirected momentum toward a set of projects that'll go public over the coming weeks. Here's what's in the pipeline.

πŸŽ₯ transformer-gimbal Β Β·Β  in progress

Natural-language-conditioned cinematic camera control on a real gimbal platform β€” Sony A7C II + DJI Ronin RS4 + ROS 2. Started with CLIP / SigLIP embeddings as a lightweight grounding layer; now pivoting to a transformer-based policy after finding the embedding-only approach under-conditioned the action distribution. Writing up the lessons as I go.

ROS 2 Β· PyTorch Β· Transformers Β· Sony SDK Β· Ronin SDK

πŸ“° eight-sprouts Β Β·Β  production AI systems

Agentic services with graph-backed retrieval and conversational reasoning, built on a Python stack with the Claude API. First product launching 2026.

FastAPI Β· PostgreSQL Β· Neo4j Β· Claude API


🧰 Working with

Machine Learning     PyTorch Β· PPO Β· GAIL Β· imitation learning Β· sim-to-real Β· evaluation design
Vision & Language    CLIP / SigLIP embeddings Β· transformer policies Β· OpenCV
Robotics             ROS 2 Β· Unity Β· PyBullet Β· Gazebo Β· gimbal & camera integration Β· sensor fusion
Engineering          Python Β· C++ Β· FastAPI Β· PostgreSQL Β· Neo4j Β· Azure Β· Docker Β· Git Β· CI/CD
Applied AI           Claude API Β· RAG Β· agentic workflows Β· knowledge graphs

πŸ“ Now

  • πŸ“ Based in Bournemouth, UK β€” open to London, European, or remote roles.
  • 🎯 Looking for research-engineer or applied-AI roles where rigorous evaluation and shipping both matter.
  • πŸ€– Especially interested in robotics foundation models, VLA systems, and AI-native products.
  • 🌱 Running Eight Sprouts on the side.
  • 🎹 Learning the piano. Slowly.

LinkedIn Β Β·Β  Scholar Β Β·Β  ORCID Β Β·Β  arXiv Β Β·Β  Email

Research + shipping. Both matter.

Pinned Loading

  1. ROS_examples ROS_examples Public

    The repository contains Basics, Motion, OpenCV, Localization, Navigation and SLAM packages in ROS.

    CMake 3

  2. DDPG-aigym DDPG-aigym Public

    Implementation of Deep Deterministic Policy Gradient Algorithm in Pytorch - Using OpenAI-Gym

    Python 2

  3. yolov5_ros1 yolov5_ros1 Public

    CMake 1