Clear, rigorous problem-solving in AI and robotics
Clear, rigorous problem-solving in AI and robotics
I built and studied a multi-robot coordination system in Webots to understand how different navigation strategies affect team performance in exploration tasks. Using Clearpath PR2 robots, I implemented and compared Route, Survey, and Mixed human-inspired strategies, along with obstacle avoidance and conflict resolution in C++. By running controlled simulations, I measured how strategy choice impacts task completion time, coverage, and coordination efficiency. One key result was that mixed strategies consistently produced more robust and efficient teams, offering insights directly relevant to real-world search, rescue, and exploration robotics.
11
175
Multi-Robot Coordination & Simulation in ROS2
Designed and implemented a ROS2-based multi-robot simulation platform to study coordination, navigation strategies, and communication-aware exploration.
Integrated SLAM Toolbox, Nav2, map merging, and multi-agent coordination logic. Used Python and C++ nodes with live visualization in RViz to evaluate efficiency, coverage, and cooperation under different strategies.🚀
21
304
Alter Ego is a personalized AI assistant I built and deployed on my website using OpenAI’s ChatGPT. It’s designed to represent how I think and work, not just answer questions. The system uses a structured persona, curated knowledge from my research and projects, and context-aware prompting to hold accurate, grounded conversations about my background in AI, robotics, and agentic systems. The goal is not imitation for novelty, but continuity: enabling clear, truthful interaction with my work, ideas, and problem-solving approach even when I’m not present.
Please check out the demo on my website:
https://amirmohaddesi.github.io/
23
237
DocForge AI is a document automation system that uses Large Language Models as reasoning engines, not just text generators. Instead of simply “writing” documents, the system analyzes structure, constraints, and intent to produce professional, publish-ready outputs such as resumes, SOC 2 reports, technical specifications, and compliance documents. LLMs are orchestrated to plan sections, enforce schemas, validate consistency, and adapt tone and layout to industry standards. The result is documents that are not only well-written, but correct, coherent, and visually polished, reducing manual effort while maintaining trust and professional quality.
Curious about how it all fits together in a connected world
Know someone who needs a haircut? 🙋
Send them to TrimsByTracy.com (http://TrimsByTracy.com?ref=drake), where they will find the salon of Tracy Turnblad (https://hairspray.fandom.com/wiki/Tracy_Turnblad), a solo hairstylist in Baltimore who needed an expert to tame her booking flow the way she tames manes.
I had a lot of fun working on this Proof of Concept, from ideating design system inspiration to working through the nooks and crannies of a hairstylist's booking flow, to figuring out how to succinctly communicate the value added in an overview video.
I also continue to learn each time I build something new with Lovable. This time I was focused on reducing the amount of time spent testing, which is what my "Process" video is about. I hope you'll take a look at that since I personally think what I shared there is very valuable and should be standard to all builders' workflow and possibly considered for productization by Lovable.
Regardless, I hope you enjoy taking a look around TrimsByTracy! All built with care in Lovable.
- Drake
https://trimsbytracy.com/
(https://trimsbytracy.com/?ref=drake)https://trims-by-tracy.lovable.app (https://trims-by-tracy.lovable.app?ref=drake)
7
3
410
Fairweather is a roadtrip planning app built around the forecast
0
52
Deficit is a platform for weight-loss coaching and community
0
61
OpenRecipe is a platform for Creators and Creatives to share, learn, and create in the kitchen.
AI Interview Copilot is a RAG-powered interview assistant that retrieves relevant information from a personalized knowledge base and uses an LLM to generate context-aware responses to interview questions. I built the retrieval and generation pipeline to combine semantic search with prompt-based reasoning, allowing responses to stay grounded in relevant source content rather than relying only on the model’s general knowledge.
0
43
DramaClips is a full-stack short-drama discovery and affiliate platform I designed, built, and deployed end-to-end. It uses Next.js, Supabase, Cloudflare R2, Railway, and Vercel, with AI integrations for content processing and automated workflows. I also built server-side tracking and attribution to measure user engagement and conversions, turning the project into a live, revenue-generating product.
0
58
This portfolio is an interactive showcase of my work across software engineering, applied AI, 3D, and creative technology. It brings together production-ready products, technical case studies, and design projects to demonstrate how I transform complex creative challenges into practical digital experiences and connected workflows.
0
50
Home3D is an AI-assisted property reconstruction application that converts residential floor plans and listing photos into an evidence-backed, interactive 3D house. Instead of asking AI to invent an entire model, it extracts room dimensions and spatial relationships into structured scene data, validates the geometry, and procedurally generates inspectable walls, floors, rooms, and supporting property evidence.