I am a Data Scientist at Lyft, working on machine learning algorithms within the Market Signals team.
In 2021, I obtained my PhD from the University of Michigan with a minor in data science. My research was primarily focused on data-driven sustainability assessment of electric vehicles, shared mobility, and vehicle automation to understand how they impact travel pattern, energy use, and economics of mobility. My work has been featured in media outlets such as PBS News Hour, E&E News, Consumer Affairs and CleanTechnica and received several prestigious awards, including Dow Sustainability Fellowship, 3M Outstanding Achievement Award, and Towner Prize.
Technical Skills:
- Causal Inference
- Machine Learning
- Transportation Economics
-
Joint PhD in Sustainability and Data Science, 2021
University of Michigan
-
MSc in Mechanical Engineering, 2015
University of British Columbia
-
BSc in Mechanical & Automotive Engineering, 2013
Iran University of Science & Technology
Experience
Recent Employments
- Developed a causal inference model to measure the effect of EV conversion on DVR productivity and LTV.
- Developed an Uplift model for EV Outreach Program, targeting +600K drivers in high-yield markets and built a personalized targeting model based on driver characteristics and travel pattern.
- Contributed to experiment design and analysis of a new incentive mechanism for retention in extreme undersupplied markets and applied robust estimation tools to de-bias marketplace interference.
- Communicated findings with the executive leadership team, presented to XFNC stakeholders, and published the results as a scientific article in a top-tier journal.
- Developed robust analytics and inference tools for Lyft’s commitment to 100% EV by 2030.
- Segmented cohorts of drivers using unsupervised learning techniques and developed EV suitability and economic metrics across individual and cohorts of drivers on the platform.
- Designed modeling framework and data pipelines for charging propensity of XD EVs, performed causal inference on time-series data from charging events, and presented actionable insights for growth and spending.
- Estimated rebound effect and induced demand of CAVs using econometric models and large survey data.
- Developed multiple machine learning, CNN, and RNN deep learning training pipelines in Python for prediction of pooling behavior using large datasets from ridesourcing trips. Models automatically retrieve data from server, optimize the training procedure on large imbalanced class datasets and perform hyperparameter tuning using grid search, achieving 94% recall and 97% precision.
- Performed causal inference (diff-in-diff and triple differences models) for robust estimation of congestion policy effect on ridesourcing pooling behavior using data from +70M trips in Chicago.
- Supervised and mentored four graduate students for data-driven and analytics research projects on mobility.
- Published +10 peer-reviewed articles in top-tier journals in transportation and sustainability research fields, presented research findings to public and expert audience, and published op-eds and blog posts.
- Acquired +$400K funding from public and private institutions including NSF and DiDi Chuxing.
- Evaluated various energy efficiency measures for electric distribution network in the city of North Vancouver.
- Identified efficiency opportunities in multi-unit commercial buildings, resulting in $550K annual energy savings.
- Developed techno-economic models of low/zero-carbon transportation infrastructures in British Columbia.
- Quantified charging infrastructure demand for deployment of battery and fuel cell electric light vehicle fleet.
- Reported and presented the results in collaboration with Pacific Institute for Climate Solution.
Awards
Selected Awards, Honors & Achievements
Publications
Media
Selected Media Coverage of Research
