nikhil ravi

Ph.D. student at Cornell University, immersed in the realm of privacy-preserving optimization algorithms for networked dynamic systems. My research primarily revolves around enhancing the robustness and privacy aspects of energy systems, with a focus on developing innovative solutions.

Currently seeking opportunities in quantitative research and data ML science for Spring 2024, I am excited about the prospect of applying my skills to impactful projects at the intersection of academia and industry.

my skills and tools

Python

NumPy

pandas

TensorFlow

PyTorch

OpenAi Gym

Ray RLlib

scikit-learn

statsmodels

plotly

Google Cloud Platform

AWS

MATLAB

R

Javascript

React.js

Next.js

Neo4J

Git

PostgreSQL

MongoDB

my education

Cornell University

Ph.D. in Electrical Engineering, Applied Mathematics and Computer Science, 2017-present

Cornell University

M.S. in Electrical Engineering, 2017-2023

Arizona State University

Ph.D. in Electrical Engineering, transferred with an M.S. to Cornell University, 2017-2021

PES Institute of Technology

B.E. in Electronics and Communications Engineering, 2013-2017

my experience

Cornell University

  • Teaching Assistant for ORIE 5380: Optimization Methods, Aug 2023 Dec 2023

  • Teaching Assistant for CS 5356: Building Startup Systems, Jan 2023 May 2023

  • Teaching Assistant for ECE 5260: Graph-based Data Science for Networked Systems, Aug 2022 Dec 2022

  • Graduate Research Assistant, Aug 2021 Present

    • Played a central role in the execution of “Provable Anonymization of Grid Data for Cyberattack Detection”, a research initiative funded by the Department of Energy as the primary student researcher, collaborating closely with my advisor and LBNL.
    • Oversaw data collection, analysis, and reporting, generating critical findings that significantly advanced the project's objectives.
    • Led innovative research initiatives within the project, including the development of
      • Differential private (DP) clustering algorithm for consumer classification and typical load shape generation
      • DP cyber-physical attack detection methodology for SCADA systems, and
      • DP approach for inferring solar PV metadata and forecasting from large-scale consumer energy usage datasets.

Kevala, Inc.

  • Data Science Intern, May 2021 Aug 2021

    • Developed a deep reinforcement learning-based tool on GCP Vertex AI to maximize batteries and plug-in electric vehicles' electricity price arbitrage value via charge schedule optimization, based on electricity price, solar irradiation, and load forecasts.
    • Built a pipeline to ingest day-ahead and real-time market electricity prices into Google BigQuery.
    • Developed a methodology to estimate carbon social prices for feeder-level electricity generation.
    • Created data visualization dashboards using Streamlit, translating complex data sets into comprehensive visual representations.
    • Researched and published an internal blog on the use of racial features in BESS adoption models.

Lawrence Berkeley National Laboratory

  • Research Intern, May 2020 Aug 2020

    • Developed a pipeline to ingest and clean large time-series AMI data of an electric ISO's consumers onto a PostgreSQL database.
    • Designed algorithms to publish differentially private summary statistics about consumer energy data.
    • Proposed a differentially private clustering algorithm to classify consumers and generate typical average load shapes of houses.
    • Developed a differential privatized cyber-physical attack detection methodology for SCADA systems.

Arizona State University

  • Graduate Research Assistant, Aug 2017 Aug 2021

    • Designed the Electron Volt Exchange, a secure Hyperledger Fabric-based distributed ledger for Transactive Energy.
    • Proposed a distributed optimization algorithm to verify users' compliance with power schedules and to mitigate the impact of false data injection.
    • Developed gradient-based edge-cutting mechanisms to build Byzantine fault-tolerant decentralized optimization algorithms.
    • Designed an algorithm to infer socioeconomic preference from crowd movement dynamics data.
    • Managed the SINE Lab's compute resource cluster including VM management, networking, and administration.