\\robin's
\\portfolio
Graduated from UCONN with a B.S.E in Computer Science with a concentration in Computational Data Analytics and a Minor in Mathematics @ UConn. 3.855 GPA, Cum Laude!! :DD
Modding thinkpads/cpus, physics and materials science, music production, Chinese language, beekeeping, trail running
Proficient: Python, Java/Kotlin, C# and .NET, C++, C // Familiar/Used for projects: Racket/Lisp/Scheme, Rust, RISC-V, Typescript/Javascript, React, CSS/Tailwind, HTML
Pytorch, Tensorflow, Keras, AWS, Apache Spark, Google Cloud Platform, Apache Hive and Pig, Docker, MySQL, Milvus, OracleDB, NoSQL/DynamoDB, Resilient Distributed Datasets (RDD), OpenCV
Transformers, LLMs, NLP, Deep Learning, RNN, CNN, Computer Vision, Reinforcement Learning
Neovim, GNU/Linux (i3 + Arch on Librebooted T400), Source Control (Git/Github), Jupyter Notebook
Driven, Analytical, Planner, Organized, Creative Problem-Solver, Teammate, Active Listener, Adaptable
The crab is from when I tried Inverse Kinematics for fun. I just really like this penguin video! (please keep scrolling ty)
Spearheading project to inform policymakers about urban electric scooter usage. Built and deployed LSTM and CNN models on AWS with Docker for usage prediction with Python, Pytorch. Data storage with Amazon RDS and inferences via AWS Lambda
Selected by professor to TA after performing best in class out of 30 students Assisted 24 undergrad computer science students with learning and understanding C++ and object oriented programming
Designed and deployed product auditing system to reduce audit time from O(n^3) to O(n) with Python. Improved and optimized auditing reports via email API, Python.
(The project on the oscilloscope) Built and designed an object-oriented painting desktop application with undo/redo and layers in C++
Optimized video content creation for livestreamer clients, impacting audience of over 870k with Python, Rest APIs, AWS. Tuned and deployed large language models on AWS Lambda on IRC data for sentiment analysis. Reduced content search times by 60+%, enabling higher throughput for increased profits
Toy model of a Python app connecting cloud hosted MySQL, MongoDB, and Milvus (a vector DB) databases to query articles with text embeddings from a sentence transformer about whether NBA players are vegetarian, and other fun stats. Yes this is an invitation to talk to me about basketball. *Editors note, Robin knows nothing about basketball
Researching novel technique for global collision detection in physics simulations, Python/C++, Building scene as vector space sum of sines via FFT, scaled Jacobian is then summed over momentum space. O(nlogn)
Fine-tuned and deployed DeTr on trail camera feeds with Pytorch for Natural Resources and the Environment students' research
Basically just my resume, wanted to make it for fun! Written in plain javascript with threejs for the 3d stuff. All models except CRT were made by me in blender. Some shaders. Deployed on vercel. Used to double as auth-flow for my Cloud Video Aggregation service. (Horrible idea, but it worked)