Mahek Vora

Experience

Where I've worked

I've been fortunate to learn from some amazing teams in quant research, consulting and industry ML, building everything from multi-agent AI tools to risk models and computer vision pipelines!

  1. JPMorgan Chase & Co.

    Jun 2024 – Jul 2026 · Bengaluru, India
    Quantitative Research Analyst
    • Architected a multi-agent RAG pipeline over a Neo4j graph and vector embeddings for ranked information retrieval and document QnA over model implementations, versions and meeting minutes, cutting months of manual review.
    • Engineered a FastAPI web app that automates building query agents over kdb+ and AWS Data Lake trading databases with multi-agent AI, adopted across Trading and Coverage teams in Global Markets.
    • Built agent evaluation with LangSmith to validate AI-generated output for correctness and version-aware filtering.

  2. JPMorgan Chase & Co.

    May 2023 – Jul 2023 · Bengaluru, India
    Quantitative Research Intern
    • Cut runtime by 12× by redesigning capital-calculation pipelines for stressed risk models (VaR, Expected Shortfall, Monte Carlo) with a sliding-window algorithm and parallelised computation.
    • Diagnosed data-quality gaps in risk-factor time series and engineered proxy-based solutions, reducing margin capital.

  3. KPMG

    Apr 2022 – Jun 2022 · Mumbai, India
    Machine Learning Intern
    • Standardised ML experiments and improved interpretability and explainability by integrating MLflow for model tracking and versioning across client engagements.

  4. Caliche Private Limited

    Dec 2021 – Feb 2022 · Guwahati, India
    Machine Learning Intern
    • Designed a two-stage UNet computer-vision pipeline in PyTorch to segment and classify defects on steel sheets, with a GUI that highlights detected defects.
    • Achieved 94% validation accuracy using residual networks, ReLU activation and AdamW optimisation. See the code.
    Steel defect segmentation outputs