Emir Ceyani
Research Assistant at USC Viterbi School of Engineering
Los Angeles, California, United States
Summary
🤩
Rockstar🎓
Top SchoolEmir Ceyani is a PhD candidate and research engineer specializing in privacy-preserving, uncertainty-aware machine learning for scientific and engineering domains, with nine years of experience across academia and industry. He builds scalable federated and distributed systems for Graph Neural Networks and Generative Flow Networks, contributing backend and gRPC integrations to the widely used FedML library. At USC Viterbi he developed agentic AI pipelines combining GNNs and LLMs for automated circuit topology and layout optimization and advanced LLM trustworthiness via conformal prediction. His work spans federated graph learning, link-level differential privacy, and diversity-seeking reinforcement learning for test-time reasoning—often targeting non-Euclidean data in high-stakes applications like MRI synthesis. A Qualcomm Innovation Fellowship finalist and NeurIPS Top Reviewer, he blends rigorous theoretical research with production-ready engineering and open-source contributions. He is seeking postdoctoral or research scientist roles starting Summer/Fall 2026 to apply generative AI and privacy-preserving ML to complex scientific challenges.
9 years of coding experience
2 years of employment as a software developer
Doctor of Philosophy - PhD, Electrical and Electronics Engineering, Doctor of Philosophy - PhD, Electrical and Electronics Engineering at University of Southern CaliforniaMaster's degree, Electrical and Electronics Engineering, Master's degree, Electrical and Electronics Engineering at Bilkent University
Stackoverflow
251reputation
7kreached
7answers
0questions
Github Skills (18)
distributed-training10
python10
mpi10
federated-learning10
grpc10
develop9
backend9
mlops8
computer-vision6
convolution6
neural-network6
keras6
lstm6
correlation6
machine-learning6
Programming languages (5)
Github contributions (5)
FedML-AI/FedML
Mar 2021 - Jul 2022
FEDML - The unified and scalable ML library for large-scale distributed training, model serving, and federated learning. FEDML Launch, a cross-cloud scheduler, further enables running any AI jobs on any GPU cloud or on-premise cluster. Built on this library, TensorOpera AI (https://TensorOpera.ai) is your generative AI platform at scale.
Role in this project:
Backend & DevOps Engineer
Backend & DevOps EngineerContributions:158 commits, 6 PRs, 92 pushes in 1 year 4 months
Contributions summary:Emir primarily focused on backend modifications and integrating gRPC communication within the FedML framework. They updated server and client backend options for federated averaging (FedAvg), modifying the code to accommodate different communication types, including MPI and gRPC. Further contributions included adapting client and server managers for the FedAvg algorithm. Additionally, the user implemented gRPC message length options and enabled gRPC communication within the FedAvg setup.
inference-enginepythonsimulationdecentralizedsilo
FedML-AI/SpreadGNN
Jun 2021 - Aug 2022
SpreadGNN: Serverless Multi-Task Learning Framework for Graph Neural Networks. Accepted to AAAI22.
Contributions:14 commits, 2 PRs in 1 year 2 months
pytorchmulti-taskdeep-learningserverlessneural-graph
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.