About

I am a PhD candidate in computer science at Cornell University, also minoring in political thought. I work on systems, and I am advised by Lorenzo Alvisi and Robbert Van Renesse. You can find me playing and making music, climbing rocks, shooting basketballs, and hanging out with friends on the lovely evenings of Ithaca.

Research Interests

My research projects address the age-old question of computer systems: how can we build reliable and efficient systems? I always look for first principles, and pursue theory and practice in tandem; a sound theoretical basis leads to principled design of practical tools and systems, which in turn inform and enrich the theoretical basis by generating insightful empirical observations.

I am currently working on metastable failures, wherein a trigger pushes a system into a self-sustaining bad mode of behavior. Metastable failures have been occurring more and more frequently, and yet are still obscure; we do not have a fundamental understanding of how they arise and why they are very difficult to detect, diagnose, and avoid. I am developing a sound and simple characterization of these failures. Such a basis is immaterial for identifying key design and operation principles, which are crucial for large-scale compute providers experiencing metastable failures. I am also building tools, including a domain-specific language called Nimbus, to reproduce empirical incidents of metastability in a controlled environment, and to look for metastability using a simple yet rich set of abstractions. This project is a collaboration with Robbert Van Renesse, Lorenzo Alvisi, Andreas Haeberlen, and Shir Cohen Gahtan.

I also pursue research on developing fast and secure consensus protocols for the permissionless model of Bitcoin. Permissionless systems have posed decentralized computing as a feasible paradigm complementing the current centralized status quo. They suffer, however, from shortcomings in performance and security. This stems from coupling performance to security. Following recent advances, we decouple them by providing deterministic safety, using the traditional quorum intersection methods. Albeit still experimental, our protocols promise further avenues to pursue provably high performance, ultimately paving the way for a practical, safe, and fast ordering layer for decentralized computing. This project is a collaboration with Youer Pu, Giuliano Losa, Ittay Eyal, and Lorenzo Alvisi.

Publications

Gorilla: Safe Permissionless Byzantine Consensus, Youer Pu, Ali Farahbakhsh, Lorenzo Alvisi, Ittay Eyal, DISC 2023.
Opportunistic routing in quantum networks, Ali Farahbakhsh and Chen Feng, IEEE INFOCOM 2022.
Dissecting the performance of chained-bft, Fangyu Gai, Ali Farahbakhsh, Jianyu Niu, Chen Feng, Ivan Beschastnikh, and Hao Duan, IEEE ICDCS 2021.

Teaching

CS 4410 - Operating Systems, Lecturer (Summer 2025, Cornell Univeristy).
CS 5414 - Distributed Computing Principles, Head TA (Fall 2024, Cornell University).
Deep and Reinforced Learning for Engineers, Lab Assistant (Spring 2021, University of British Columbia).
Stochastic Analysis, TA (Winter 2019, Sharif University of Technology).
Data Communication Networks, TA (Winter 2018, Sharif University of Technology).

''We are what we repeatedly do. Excellence, then, is not an act, but a habit.'', Aristotle.