DOCX vCard Plain text

Dr. Priya Sharma

Senior Data Scientist

I build statistical models that change how organizations make decisions. PhD in computational neuroscience, six years in industry turning messy data into defensible conclusions.

Boston, MA priyasharma.io Google ScholarGitHubLinkedIn

About

I spent four years in a neuroscience lab studying how the brain encodes uncertainty, then realized the same math applies to every business problem I've seen since. At Moderna I built the forecasting models that guided $200M in manufacturing capacity planning. At Two Sigma I developed causal inference pipelines that reshaped how portfolio managers think about factor attribution.

I care about reproducibility, honest uncertainty quantification, and making sure the person reading my analysis can actually act on it. I publish in peer-reviewed venues and ship production ML systems — I don't think those should be separate careers.

Experience

Senior Data Scientist, Supply Chain Intelligence

Moderna

2023 – Present Cambridge, MA

Lead data scientist for demand forecasting and manufacturing optimization across the mRNA vaccine portfolio.

  • Built hierarchical Bayesian demand model covering 14 markets; reduced forecast MAPE from 32% to 11%, directly informing $200M+ in capacity investment.
  • Designed causal impact framework for evaluating process changes on yield; identified two interventions worth $18M/year in recovered product.
  • Created an internal Python library (moderna-ts) for time-series preprocessing now used by 40+ analysts across three business units.
  • Mentored two junior scientists through their first production model deployments.

PythonPyMCStanSnowflakeAirflowDocker

Data Scientist, Factor Research

Two Sigma

2020 – 2023 New York, NY

Quantitative research on causal factor models for systematic equity strategies.

  • Developed a double-ML causal attribution pipeline that decomposed portfolio returns into 23 interpretable factors with confidence intervals; adopted firm-wide.
  • Built real-time anomaly detection for data vendor feeds using conformal prediction; caught 14 data quality incidents before they reached production.
  • Published internal research note on synthetic control methods for strategy evaluation, later adapted into a NeurIPS workshop paper.

PythonRSparkKubernetesPostgreSQL

Postdoctoral Researcher

MIT — McGovern Institute

2018 – 2020 Cambridge, MA

Computational neuroscience research on probabilistic models of sensory decision-making.

  • First-authored paper in Nature Neuroscience on Bayesian models of multi-sensory integration in the primate cortex (87 citations).
  • Built an open-source Julia package (BayesBrain.jl) for neural population decoding; 600+ GitHub stars.
  • Co-taught MIT 9.40 (Introduction to Neural Computation) for two semesters.

JuliaPythonMATLABStanSLURM

Selected Publications

  1. Sharma P, Chen L, Bhatt R

    Bayesian multi-sensory integration in primate parietal cortex

    Nature Neuroscience 2019

    Demonstrated that neurons in area MSTd implement near-optimal Bayesian cue combination, with reliability-weighted priors learned from experience.

  2. Sharma P, Goldstein A

    Causal factor attribution with double machine learning: a practitioner's guide

    NeurIPS Workshop on Causal Inference 2022

    Practical framework for applying DML to factor models in finance, with finite-sample corrections for high-dimensional confounders.

  3. Sharma P, Liu W, Park J

    Conformal anomaly detection for streaming financial data

    ICML 2023

    Real-time anomaly detection with finite-sample coverage guarantees, evaluated on three years of equity market data.

  4. Sharma P, Rodriguez M

    Hierarchical demand forecasting for pharmaceutical manufacturing

    arXiv preprint (under review at Management Science) 2025

    Bayesian hierarchical model for coherent demand forecasts across product, geography, and time hierarchies.

Open-Source Projects

Technical Skills

Statistical Methods

Bayesian inferenceCausal inferenceTime seriesSurvival analysisExperimental design

ML & Deep Learning

PyTorchscikit-learnXGBoostTransformersConformal prediction

Languages

PythonRJuliaSQLStan

Infrastructure

AirflowSparkDockerKubernetesAWSSnowflake

Education

Awards & Honors

Certifications

Languages

Community

Mentor

Data Science for Social Good

2021 – Present

Summer fellowship mentor; guided three teams building ML models for non-profits in public health and criminal justice reform.