Open to opportunities

Poorva Kemse

@poorvakemse

Data Scientist building Bayesian state-space and mixture-regime models for structural signal extraction and tail-risk forecasting.

India

Message

What I'm looking for

I’m looking to apply Bayesian modeling to quantify structural risk and improve model integrity—working on stochastic regime analysis, tail-risk forecasting, and decision-support analytics in data-rich, high-noise environments.

I build Bayesian systems that turn noisy, non-stationary signals into interpretable structure—starting with a Global Coffee Demand Forecast 2026 using a Bayesian NDLM in R, tuned via discounting to reduce out-of-sample RMSE (7.77) for stockout risk mitigation.

Most recently, I’m running a NorCal seismic reliability audit (ETAS + Bayesian state-space) and engineering a Gamma-Exponential-Lomax mixture to separate seismic regimes and quantify mechanical energy loss, using MCMC to keep the model’s integrity under heavy-tailed uncertainty.

Experience

Work history, roles, and key accomplishments

Sr. Research and Development Engineer

Indkarta

Feb 2021 - Dec 2021 (10 months)

Co-inventor of 'Agestek' (Patent: IN 401064), an IoT bio-security framework calibrated to CDC standards. Developed multi-sensor fusion and RFID-to-Cloud pipelines.

IoT Sensor Fusion Embedded Systems

Embedded System Intern

Exa Mobility

Mar 2020 - Jun 2020 (3 months)

Benchmarked Apple A13/A14 and NVIDIA Jetson architectures for edge-computing. Engineered IMU signal processing for real-time motion tracking.

Embedded Systems Edge Computing IMU Signal Processing

Electronics Team Head

PICT Robotics

Aug 2016 - Aug 2019 (3 years)

Designed Master-Slave control systems and high-current MOSFET drivers with surge prevention. Engineered omni-directional Swerve Drive with integrated LSA for automated path-correction.

Control Systems MOSFET Swerve Drive