Mobirise

Achraf Bahamou

Ph.D Candidate, Deming Center Doctoral Fellow.
Industrial Engineering and Operations Research Department, Columbia University.

I am a final year PhD Candidate in
Operations Research at Columbia University.

My research interests include:
- Data-driven decision-making, Value of information and mechanism design in low/finite informational environments.
- Optimization algorithms for Machine learning and Deep learning.

I am fortunate to be advised by Prof. Donald Goldfarb and Prof. Omar Besbes.

Prior to joining Columbia, I graduated from
Ecole Polytechnique, Paris with BSc and MSc degrees in Applied Mathematics and Statistics. 

Linkedin

C.V 

E-mail:

achraf.bahamou [at] columbia [dot] edu

Google Scholar

Industry Experience

Mobirise

Quantitative Research Intern, Jump Trading 

Chicago, IL

- Research, feature engineering, model selection and writing execution algorithms.

Mobirise

Applied Scientist Intern, Amazon

Seattle, WA

- Research on Data-Driven Pricing using Machine Learning.

Mobirise

Data Scientist Intern, Google 

New York, NY

- Worked with Active-Learning-For-All (ALFA) and Neurosurgeon teams, Google Research NYC.
- Research on active learning gains forecasting.

Mobirise

Quantitative Research Intern, Hellebore Capital Limited 

London, UK

- Worked on modeling irregularly spaced trades arrival times using multivariate Hawkes processes.
- Built proprietary python packages to deal with data extraction, preprocessing, fit and simulate different Hawkes Processes models. Published a research paper in ITISE 2018 conference.

Mobirise

Data Analyst Intern, Air France-KLM 

Paris, FR

Revenue Management and Decision Support Team
- Worked on estimating the unconstrained demand on long-haul flights using time series forecasting models and machine learning.

Research

My research in data-driven decision-making focuses on characterizing the value of information and optimal pricing mechanisms in low/finite informational environments:

- Optimal Pricing with a Single Point. joint with Prof.Omar Besbes and Amine Allouah

Accepted in EC 2021, Major Revision in Management Science [arxiv.org] [Informs 2020 Talk].

Presented in RMP 2021 (spotlight session), MSOM 2021, EC 2021, Informs 2020

- Pricing with Samples. joint with Prof.Omar Besbes and Amine Allouah. 

Accepted in EC 2021, Operations Research (forthcoming). [SSRN]

- Optimal Limited Price Experimentation. joint with Prof.Omar Besbes and Amine Allouah

(Ongoing work).


My research in Optimization involves developing efficient 1st and 2nd order optimization algorithms for minimizing loss functions in Machine Learning and Deep Learning frameworks:

- Practical BFGS Methods for Training Deep Neural Networksjoint with Prof. Goldfarb and Yi Ren 

Accepted in Neurips 2020, Spotlight presentation [arxiv.org] 

- A Mini-Block Natural Gradient Method for Deep Neural Networks. joint with Prof. Goldfarb and Yi Ren

Submitted [arxiv.org] 

- Kronecker-factored Quasi-Newton Methods for Deep Learning. joint with Prof. Goldfarb and Yi Ren

Submitted [arxiv.org] 

- An Adaptive Step-Size Procedure for Stochastic First-Order Optimization Methods.joint with Prof. Donald Goldfarb (Ongoing work).

- Stochastic Flows and Geometric Optimization on the Orthogonal Group.joint with Krzysztof Choromanski, Google Brain NY et al. Accepted in ICML 2020 [arxiv.org] [ICML 2020 Proceedings]


Previous research:

- Hawkes processes for credit indices time series analysis: How random are trades arrival times? 

joint with Maud Doumergue, Philippe Donnat, Hellebore Capital LLC. ITISE 2018 International Conference on Time Series and Forecasting accepted paper[arxiv.org] [ITISE 2018 Proceedings]

Education

Mobirise

Columbia University in the City of New York

2018-Present

Ph.D. Candidate in Operations Research. Data Science track.
- Advisors: Prof. Donald Goldfarb and Prof. Omar Besbes.
- Deming Center Doctoral Fellow.
- Relevant Courses: Machine Learning and High-Dimensional Data, Optimization I and II, Stochastic Modeling I and II, Reinforcement Learning, Theoretical Statistics I, Quantum Computing.

Mobirise

Ecole Polytechnique, Paris

2015-2018

Bachelor of Science and Master of Science in Applied Mathematics.
- Relevant Courses: Machine learning, Statistics, Times Series Analysis, Stochastic processes and Monte-Carlo methods, Optimization, Random modeling, Analysis of Algorithms, Statistical Physics, Quantum Mechanics.

Awards


- Deming Center Doctoral Fellowship. Deming Center Doctoral Fellow 2021.

- Postgraduate Excellence Scholarship. OCP Foundation.

- French Government Major-Excellence Scholarship. The French Ministry for Foreign Affairs.

- Hassan II Academy of Science and Technology Fellowship. Ranked First in The National Open Competition Of Science and Technology - Physics category.

Teaching


- Optimization Models and Methods, M.S OR core course (IEORE4004), Columbia University

Head Teaching Assistant (Spring 2021), Around 170 students

- Columbia Center for Teaching and Learning, Research Assistant (Fall 2019, Spring 2020)

Collaborating with the Center for Teaching and Learning to introduce an auto-grading framework and incorporate digital technology into the teaching environment to facilitate active learning.  

- Business Analytics, (IEOR 4650), Columbia University

Head Teaching Assistant (Fall 2019) , Around 70 students [Info]  

- Simulation Modeling and Analysis, (IEORE3404), Columbia University

Head Teaching Assistant (Spring 2019, 2020), Around 90 students [Info]  

- Simulation, M.S OR core course (IEORE4404), Columbia University

Head Teaching Assistant (Fall 2018, 2020), Around 130 students [Info] 

Contact 

Email: achraf.bahamou [at] columbia [dot] edu
Office Address:  Achraf Bahamou     
                  500 West 120th Street, Mudd 321  
                  Columbia University, New York 10027, USA

© Achraf Bahamou 2022

Built with Mobirise web themes