Portfolio

Projects & Research

A collection of my work in AI/ML, robotics, and autonomous systems. From production LLM applications to cutting-edge research publications.

AI and Machine Learning Projects

GitHub Repositories

Featured Projects

Open-source tools and frameworks I've built for the AI/ML community

Agentic LLM Framework

A modular framework for building autonomous LLM agents with planning, memory, and tool-use capabilities. 1,250+ GitHub stars.

ROS2 Deep RL Navigation

Deep reinforcement learning-based autonomous navigation for mobile robots using ROS2. Implements PPO and SAC with sim-to-real transfer.

Multimodal RAG System

Production-ready RAG system supporting text, images, and code with hybrid search, reranking, and streaming responses.

Robotic Manipulation Transformer

Transformer-based policy for robotic manipulation tasks with strong generalization to novel objects.

LLM Code Reviewer

Automated code review tool powered by LLMs. Analyzes PRs for bugs, security issues, and style violations.

Vision-Language Robotics

Vision-language models for robotic task understanding. Enables natural language instruction following.

Research

Selected Publications

8 Publications

Published at top-tier venues including NeurIPS, ICRA, ACL, EMNLP, CoRL, and AAAI.

Multi-Agent Coordination in LLM Systems (NeurIPS 2025)

A hierarchical planning framework for coordinating multiple LLM agents, achieving 40% improvement in task completion rates.

Sim-to-Real Transfer for Robot Navigation (ICRA 2025)

Domain randomization strategy enabling zero-shot transfer of navigation policies with 92% success rate.

Efficient Fine-Tuning of LLMs (ACL 2024)

Parameter-efficient fine-tuning reducing memory by 70% while maintaining 98% performance.

Vision-Language Models for Robotics (RA-L 2024)

Comprehensive survey and RoboVLM-Bench benchmark for evaluating VLM-based robot policies.

Real-world impact through production deployments.

Production LLM Pipeline

Built and deployed an LLM pipeline serving 10K+ daily requests at Melbourne AI Lab.

SLAM Navigation System

Implemented SLAM-based navigation reducing localization error by 35% for mobile robots.

Deep RL Manipulation

Designed deep RL policy for robotic arm manipulation achieving 95% success rate.

Computer Vision QA System

Developed computer vision system for quality inspection with 98% accuracy.

By the Numbers

8
Research Papers
2.5K+
GitHub Stars
25+
Projects Completed
12
Conference Talks

Interested in Collaboration?

I'm always open to discussing new projects, research collaborations, or consulting opportunities.