Profile picture
Jason Zafirov, PhDSenior AI Quant
Education
PhD, Anderson School of Management, University of California, Los Angeles (UCLA)
  • •Focus: AI/​ML Applications in Finance
  • •Thesis: Improving Inference with Machine Learning: Application to CEO Turnover⁠
    • 2018 – 2023Los Angeles, CA
    Master's of Science, Applied Economics, HEC Montréal
  • •Master's Thesis and Focus: Macroeconomics⁠
    • 2012 – 2015Montreal, QC
    Bachelor's of Commerce, Finance, John Molson School of Business, Concordia University
  • •Focus: Portfolio Management (Calvin Potter Fellow: member of The Kenneth Woods Portfolio Management Program, hands-on training in investment management of an endowment fund portfolio of over $5 million⁠)
    • 2008 – 2011Montreal, QC
    Diploma of College Studies, Computer Science, Dawson College

    with semester abroad at Claude Bernard University Lyon 1 in Lyon, France

    2005 – 2008Montreal, QC
    ㅤ
    Skills
    PythonSQLRSparkGenAI / LLMsNeural NetworksEnsemble TreesRAG AgentsRAG pipelinesMCPsKnowledge Graph RAGsFinancial/Economic AnalysisFactor AnalysisBacktesting/ForecastingQuantitative AnalysisRisk ManagementAWSGithubPipeline BuildingInvestment Research Reporting
    Professional Experience
    Senior AI Quant, Pleteum Family Office (AUM: $650M)
  • •Long-term forecasting, factor rotation and multi-asset allocation strategies using ML-enhanced factor models and signals for active management of public equities.
  • •Risk Management with a particular focus on Private Equity and Venture Capital holdings.
    • May 2025 – presentDubai, UAE
  • •Built MCP with Knowledge Graph Agentic RAG that allows investment management to interface with assets, backtest investment ideas and stress test the impacts of hypothetical geopolitical and economic scenarios.
  • •Actively managed $67.5 million public equity portfolio.
  • •Performance on quarterly rebalance strategy (3 month holding periods):

  • •Toolkit: ChatGPT API, MSCI BarraOne, MSCI MAC/​GEM/​PEQ2 factor models, Preqin datasets, S&P CIQ and Private Company datasets, Snowflake, AWS, QuiverQuant.

  • •Coding languages used: OpenAI API, Claude Code, MSCI BarraOne, MSCI MAC/​GEM/​PEQ2 factor models, Preqin datasets, S&P CIQ and Private Company datasets, SEC filings, Congress Trade Filings, Snowflake, AWS.
    • AI Lead, Integrated Data Intelligence Ltd (Startup)
  • •AI applied to Venture Capital /​ Private Equity. Startup success (probability of successful exit, return and valuation) forecasting using financial relationship modeling and different AI algorithms: neural networks, gradient-boosted trees and LLM, NLP methods (including embeddings and text chunking) with GenAI & RAG pipelines.
  • •Toolkit: Palantir Foundry, ChatGPT API, Crunchbase, PitchBook
    • Apr 2023 – Apr 2025Abu Dhabi, UAE
  • •Coding Languages: Python, PySpark, SQL
    • Data Scientist, EPAM (NYSE: EPAM, MktCap: $20B)
  • •Time series forecasting investor behavior using a variety of machine learning algorithms.
  • •Event categorization, article clustering and sentiment analysis through NLP.
    • Jul 2022 – Apr 2023EU - Remote
  • •Toolkit: Palantir Foundry, Databricks, HubSpot
  • •Coding Languages: Python, PySpark, R, SQL
    • Economist, Letko Brosseau (private fund, AUM: $27B)
  • •Forecasting and modeling: emerging market economics, energy markets, corporate profits, electric vehicle supply, sovereign debt.
  • •Investment recommendations writing. Data organization, presentation and visualization.
    • Jan 2014 – Sep 2016Montreal, Canada
  • •Toolkit: S&P Capital IQ, Bloomberg Terminal
    • Intern — Quantitative Analysis, CDPQ - Québec Public Pension Plans (AUM: $286B)
  • •Built PCA autoregression tool to forecast inflation, model: FAVAR based on principal components established from over 150 time series.
  • •Toolkit: Datastream
    • May 2013 – Aug 2013Montreal, Canada
  • •Coding Languages: MATLAB
    • Intern — Macroeconomic Analysis, CDPQ - Québec Public Pension Plans (AUM: $286B)
  • •Forecasting and modeling: unemployment, gross domestic product, demographics, productivity.
  • •Toolkit: Datastream
    • Jan 2013 – Apr 2013Montreal, Canada
  • •Coding Languages: Eviews
    • Research Analyst — Capital Markets, Scotiabank
  • •Portfolio Strategy team. Macroeconomic and style factor forecasting.
  • •Toolkit: Bloomberg Terminal, S&P Capital IQ.
    • May 2010 – Jul 2011Montreal, Canada
    Intern — Trading Floor, Healthcare of Ontario Pension Plan (AUM: $114B)
  • •Software development for trading floor operators.
  • •Toolkit: Datastream, S&P Capital IQ
    • May 2009 – Aug 2009Toronto, Canada
  • •Coding languages: VBA
    • Intern — Software Development, CEDIB
  • •Coding languages: Adélia (which is a French procedural programming language)

    • Apr 2008 – Jun 2008Lyon, France

    Publications — Authorships
    A Comprehensive Look at the Empirical Performance of Equity Premium Prediction II⁠, The Review of Financial Studies (Top 5 Journal in field)
    Private Market Portfolio Forecasting and Optimization Using Machine Learning, United States Patent and Trademark Office (Provisional Patent)
    Publications — Research Assistance
  • •Analysis of minimum wage changes on store and employee outcomes of large US retailer (~5,000 stores). Linking together HR data, performance, turnover, scheduling, payments, geography, revenues and web scraping variables.
    • Government Debt Management: The Long and the Short of It⁠, Published, The Review of Economic Studies
  • •Statistical analysis of US debt dynamics.
  • •Researched conditions of unemployment benefits.
    • Household Search and the Aggregate Labor Market⁠, Published, The Review of Economic Studies
  • •Statistical analysis using US CPS household data.
    • Counterproductive Worker Behavior After a Pay Cut⁠, Published, Journal of the European Economic Association
  • •Difference-in-difference regression modeling of how workers reacted to a pay cut in a sales call-center.
    • Machine Learning Projects
    pandas-ai⁠, Contributor

    Library that adds Generative AI capabilities to pandas.

    Professional Training
    S2DS, Europe’s largest data science training bootcamp
    2023London, United Kingdom

    Machine Learning PhD Classes
    COM SCI 263: Natural Language Processing, UCLA Computer Science Department, Grade: A
    ENGR C247: Neural Networks & Deep Learning, UCLA Engineering Department, Grade: A
    BIOSTAT 273: Classification and Regression Trees, UCLA Biostatistics Department, Grade: A+
    STATS 232C: Cognitive Artificial Intelligence (Reinforcement Learning), UCLA Statistics Department, Grade: A+

    Teaching Assistance
    MGMT 439: Tools and Analysis for Business Strategy (introduction to data science), Graduate level (UCLA)

    MGMT 264A: Market Research (regression, dimensionality reduction and clustering), Graduate level (UCLA)

    MGMT 403: Financial Accounting, Undergraduate level (UCLA)
    ECO 680109: Applied Econometrics, Graduate level (HEC Montréal)
    ECO 600112A: Economic & Corporate Organization, Graduate level (HEC Montréal)
    ECO 380607: Introduction to Econometrics, Undergraduate level (HEC Montréal)
    Personal Tutor, Economics, Computer Science, Data Science.
    Awards
    UCLA Anderson doctoral fellowship, $140,000
    FRQSC - Provincial doctoral scholarship, $84,000
    Academic Excellence Award (2012 & 2013)

    Highest average of the Master's program.

    Academic Excellence Award (2007)

    Highest average of the Computer Science program.

    Languages
    English

    (Native)

    French

    (Fluent)

    Spanish

    (B1)