Summary

"Welcome to my About Me page! In 2017, I graduated from the Sharif University of Technology with a Master's degree in Artificial Intelligence.

My career kickstarted at Miras Technology, a technology provider company. As a data scientist there, I had the opportunity to tackle various data science challenges. From developing recommendation engines to solving loan credit risk scoring problems, I delved into the fascinating world of data-driven solutions. Additionally, I gained valuable experience in data engineering and encountered intricate issues related to model serving while contributing to Miniature, a media monitoring platform. This hands-on involvement with big data and cutting-edge technologies like ElasticSearch, Kafka, and Redis broadened my skill set.

Recognizing my affinity for solving complex problems, I ventured into the realm of quantitative finance during a Miras project involving algorithmic trading. Eager to deepen my knowledge, I extensively studied quantitative finance and even completed the Oxford Algorithmic Trading Course, earning certifications along the way. This newfound passion led my coworkers and me to co-found Eveince, a fintech company specializing in providing technology solutions for hedge funds. As the co-founder and Quant Team Lead at Eveince, I spearheaded the designing and developing of a robust quantitative trading system, implementing advanced asset allocation strategies and risk management mechanisms. Furthermore, I was responsible for designing and managing Eveince's cryptocurrency funds, each tailored to specific risk characteristics and driven by quantitative models in spot (long-only) markets.

Drawing from my experience at Miras, I worked on a side project at Eveince that involved building a recommendation system for Fidibo, the largest online bookshop in Iran. As the Data Team Lead, I managed the team's efforts to deliver a high-performing recommendation engine.

Driven by my passion for complex problem-solving, I recently joined Digikala, the leading e-commerce platform in the Middle East, as an AI/Data Product Manager. In this role, I oversee the development and strategy of the recommendation engine product, aligning it with the company's objectives and leveraging the cutting-edge technologies crafted by the tech team.

What intrigues me about both recommendation engines and quantitative trading is their multifaceted nature. They resemble a game of poker, where statistical concepts and expected value play a vital role in crafting effective solutions. Furthermore, these fields are not solely governed by algorithms but are also influenced by various other variables. To tackle these challenges, I believe in employing comprehensive risk management mechanisms and supplementary tools. This dynamic environment fuels my passion for these fields and keeps me motivated to explore new horizons.

Resume

Working Experience

  • Quant-Team Lead and Data Scientist

    Eveince Capital | June 2020 - Present

    - Spearhead quantitative technologies and data-driven product development at Eveince, a technology provider specializing in quantitative finance.

    - Lead the design and implementation of the Eveince ML-trading pipeline, consisting of three quantitative modules.

    - Developed quantitative trading models based on market expectation estimators, bet sizing techniques, and risk objectives using the VAR method.

    - Applied deep reinforcement learning (DQN) to optimize order placement and maximize VWAP (Volume-Weighted Average Price).

    - Utilized Monte Carlo simulations and Modern Portfolio Theory (MPT) optimization for portfolio rebalancing strategies.

    - Designed and implemented the Eveince backtest framework with a layered architecture, enabling comprehensive evaluation of trading strategies.

    - Defined fund objectives and risk characteristics, contributing to the development of different funds Eveince Capital offers.

  • Data Product Manager

    Digikala | Since Jan. 2023

    - Design recommendation engine strategies to optimize the user experience and achieve business objectives.

    - Conduct data analysis using advanced data science techniques and perform exploratory data analysis (EDA) to uncover valuable insights and drive data-informed decision-making.

    - Collaborate with cross-functional teams to implement data-driven solutions and leverage data to improve various aspects of Digikala's operations.

  • Data-Team Lead and Data Scientist

    Miras Technology | May 2018 - Apr. 2020

    - Managed the data and research team, defined the technology roadmap, and assessed the team.

    - Designed data products, including social media and news trend extraction, loan credit risk scoring, advertisement recognition, and data deduplication.

    Fidibo Recommendation:

    • Worked on the recommendation engine for Fidibo, the largest online book marketplace in Iran.
    • Implemented various Recommendation tools like content-based (embedding books), Colds Starts & User-Based recommendations for different use cases and scenarios.
    • Designed a Tracker and analytic panel for evaluating recommendation models and other business flows.

    AVA:

    • Developed and implemented a systematic trading system, AVA, which automated quantitative algorithms for multi-financial markets, including the Iranian stock market and cryptocurrencies.

    Miniature:

    • Developed data augmentation pipelines and a trend extraction module for a faceted search engine. Miniature was a Media Monitoring platform that enables customers to extract insight and data from social media like Twitter, Telegram, and News.

Research Experience

  • Education

    • Master of Science in Artificial Intelligence | Sept. 2014 - Sept. 2016
      The Sharif University of Technology
      Thesis: 3D Soccer Ball Tracking with Fixed and PTZ Cameras
    • Bachelor of Science in Computer Engineering | Sept. 2010 - Sept. 2014
      Isfahan University of Technology

  • Laboratories

    • Image Processing Laboratory | Sept. 2016 - Feb. 2017
      The Sharif University of Technology
      Supervisor: Prof. Shohre Kasaei
    • Ambiance Intelligence Laboratory (AIRLAB) | Sept. 2016 - Feb. 2017
      Sharif ICT Innovation Center
      Under Supervision: Prof. Hamid Karbalaee Aghajan

  • Publication

    • Credit Risk Rating Using State Machines and Machine Learning
      B. Sabeti, H.A. Firouzajee, R. Fahmi, S.H.E. Mortazavi Najafabadi
      ICEFR, 2020
    • Irony Detection in Persian Language: A Transfer Learning Approach Using Emoji Prediction
      P. Golazizian, B. Sabeti, A. Ashrafi, Z. Majdabadi, O. Momenzadeh, R. Fahmi
      LREC, 2020
    • Twitter Trend Extraction: A Graph-based Approach for Tweet and Hashtag Ranking, Utilizing No-Hashtag Tweets
      Z. Majdabadi, B. Sabeti, P. Golazizian, A. Ashrafi, O. Momenzadeh, R. Fahmi
      LREC, 2020
    • Optimizing Annotation Effort Using Active Learning Strategies: A Sentiment Analysis Case Study in Persian
      A. Ashrafi, B. Sabeti, Z. Majdabadi, P. Golazizian, O. Momenzadeh, R. Fahmi
      LREC, 2020

  • Published Contents

    • Art of Problem Solving with Data | July 2021
      Instructor and designer of online course Art of Problem Solving with Data which is deployed on maktabkhooneh.org platform

Skills

  • Programming Languages & Frameworks: Python (Pandas, NumPy, Scikit-Learn, Keras, Plotly, OpenCV, Dlib, Flask), Scala (Akka, AkkaHTTP), C & C++ (Qt)
  • Tools & Technologies: Git, CI-CD, Docker, ElasticSearch, Grafana, Apache Kafka, Redis, Apache Airflow, MinIO S3, Unity, K8S

Languages: Persian (native), Arabic (native), English (fluent)

Other Activities & Hobbies

Hobbies: Poker, Sculpture, Theater

Sports: Tennis, Basketball