Hello, my name is
Timothy Castiglia, PhD
Machine Learning Researcher
- castigliatim@gmail.com
- 1-845-661-6818
- GitHub
About me
I am a Machine Learning Researcher with a PhD in efficient distributed artificial intelligence & years of industry experience.
I prioritize the cost-efficiency and scalability of AI systems.
Skills
- Python - Deep learning
- C / C++ - Fast low-level computation
- Matlab - Matrix operations and visualization
- Rust - Fast web backend
- Bash - Scripting
- Java - App development
- PyTorch - Custom DL algorithms
- TensorFlow - Efficient DL algorithms
- Docker - Replicable environments
- AWS EC2 - Simulating distributed systems
- Git - Version control
- Jenkins - Software testing and verification
- Software Lead and Mentor 2016-2018
- Teaching Assistant 2018-2019
- Research Mentor 2021-2023
- IBM Research Collaborator 2020-2023
- Windows
- Ubuntu
- CentOS
- MacOS
- Raspbian
- Manjaro
- GNU/Linux
Work Experience
IBM Research
Summer 2022
Research Extern
Designed novel algorithm of feature selection for deep neural networks in Vertical Federated Learning settings without sharing private features. Verified method using PyTorch and Tensorflow in IBM's Federated Learning library, demonstrated it can reduce computation and communication cost. Filed patent and first author on paper published in ICML 2023.
IBM Research
Summer 2021
Research Extern
Studied Vertical Federated Learning in collaboration with several researchers at IBM. Tackled several unsolved problems, including theoretical convergence with message compression, viability of privacy-preserving techniques, and flexibility of participant model updates. Experiments required extensive use of PyTorch and Tensorflow in HPC systems. Filed patent and first author on paper published in ICML 2022.
Black River Systems
Summer 2018
Software Engineer
Created back-end for several modules, written in Rust, for a system to monitor mission-critical equipment.
Critical Technologies
2016-2018
Lead Developer
Software lead of an online Spectrum Analyzer for analyzing interference in 4G networks. Application required fast streaming of spectra (∼ 5000 traces/second) from FPGA to disk through C++ interface, and streaming from Python backend to web front-end application. Led integration of hardware and software aspects of the product and mentored new software developers on the team.
Education
Rensselaer Polytechnic Institute
2018 - 2023
Doctor of Philosophy (PhD) in Computer Science
Rensselaer Polytechnic Institute
2013-2017
Bachelor of Science in Computer Science and Computer Engineering
Portfolio & CV
Use the template as-is or try different colors and fonts from the options below.
Try Other Colors
Try Other Fonts
- Barlow Semi Condensed/Barlow Semi Condensed
- Aa
- Aa
- Aa
- Aa
- Aa
- Aa
- Aa
- Aa