Funding
Competitively funded programs across U.S. agencies and private sector.
TOTAL AWARDS: 16 | GROSS FUNDING: ~$33.4M
U.S. Department of Energy (DOE)
2025–2027
Genesis Mission / ModCon — AI Safety & Security Thrust $30,000,000
Team Member — Office of Advanced Scientific Computing Research (ASCR)
Deliverable: Threat models, safeguards, and evaluation protocols for AI safety in DOE foundation-model deployments across national laboratory workflows.
Team Member — Office of Advanced Scientific Computing Research (ASCR)
Deliverable: Threat models, safeguards, and evaluation protocols for AI safety in DOE foundation-model deployments across national laboratory workflows.
2024–2027
Federated Learning Methods for Privacy Preservation in Foundational Models $7,000,000
Principal Investigator — Office of Advanced Scientific Computing Research (ASCR), AI for Science Project deliverables
Deliverable: Privacy-preserving federated training methods for scientific foundation models, with open-source implementations and cross-institutional pilots at DOE facilities.
Principal Investigator — Office of Advanced Scientific Computing Research (ASCR), AI for Science Project deliverables
Deliverable: Privacy-preserving federated training methods for scientific foundation models, with open-source implementations and cross-institutional pilots at DOE facilities.
2023–2025
Physics-Aware Acceleration for Federated Learning on Scientific Data $900,000
Co-Principal Investigator — Office of Advanced Scientific Computing Research (ASCR), EXPRESS Program
Deliverable: Physics-informed compression and acceleration methods that reduce federated learning communication costs for large-scale scientific simulations.
Co-Principal Investigator — Office of Advanced Scientific Computing Research (ASCR), EXPRESS Program
Deliverable: Physics-informed compression and acceleration methods that reduce federated learning communication costs for large-scale scientific simulations.
2019–2020
Scalable Load Management Using Reinforcement Learning $3,000,000
Task Lead — Office of Energy Efficiency and Renewable Energy (EERE)
Outcome: Deployed reinforcement learning controllers for adaptive grid load management, demonstrating scalable demand-response across heterogeneous building portfolios.
Task Lead — Office of Energy Efficiency and Renewable Energy (EERE)
Outcome: Deployed reinforcement learning controllers for adaptive grid load management, demonstrating scalable demand-response across heterogeneous building portfolios.
U.S. Department of Energy — National Nuclear Security Administration (NNSA)
Explainable, High Confidence Models for Dynamic Systems $3,000,000
Task Lead — DOE NNSA
Outcome: Explainability and uncertainty quantification methods for ML models on dynamic physical systems in national-security environments.
Task Lead — DOE NNSA
Outcome: Explainability and uncertainty quantification methods for ML models on dynamic physical systems in national-security environments.
Intrusion Detection for Controller Area Networks for Vehicles $1,000,000
Task Lead — DOE NNSA
Outcome: ML-based intrusion detection for in-vehicle CAN bus communications, advancing cybersecurity for autonomous and connected vehicle platforms.
Task Lead — DOE NNSA
Outcome: ML-based intrusion detection for in-vehicle CAN bus communications, advancing cybersecurity for autonomous and connected vehicle platforms.
U.S. Department of Veterans Affairs
2023–2025
Lung Cancer Prediction and Explainability
Task Lead — U.S. Veterans Affairs
Outcome: Privacy-preserving predictive models and explainability tools for lung cancer risk stratification across VA patient cohorts.
Task Lead — U.S. Veterans Affairs
Outcome: Privacy-preserving predictive models and explainability tools for lung cancer risk stratification across VA patient cohorts.
U.S. Department of Defense (DoD)
2020
Misinformation Impact on Covid-19
Task Lead — DoD
Outcome: Analytical framework assessing the propagation and societal impact of health misinformation during the COVID-19 pandemic.
Task Lead — DoD
Outcome: Analytical framework assessing the propagation and societal impact of health misinformation during the COVID-19 pandemic.
2020
Activity Characterization $200,000
Task Lead — DoD
Outcome: Behavioral pattern recognition and activity characterization methods for national-security applications.
Task Lead — DoD
Outcome: Behavioral pattern recognition and activity characterization methods for national-security applications.
ORNL Laboratory Directed Research & Development (LDRD / SEED)
2024–2025
Scalable Privacy Algorithms and Benchmarks for Federated Learning Frameworks (PRESTO) $240,000
Principal Investigator — ORNL LDRD
Deliverable: PRESTO — a reproducible benchmark suite for evaluating differential privacy, secure aggregation, and gradient protection in federated scientific learning.
Principal Investigator — ORNL LDRD
Deliverable: PRESTO — a reproducible benchmark suite for evaluating differential privacy, secure aggregation, and gradient protection in federated scientific learning.
2021–2023
Automation for Grid Interconnected-Laboratory Emulation $900,000
ML Advisor — ORNL LDRD
Outcome: Machine learning advisory on automated emulation environments for grid interconnection testing at laboratory scale.
ML Advisor — ORNL LDRD
Outcome: Machine learning advisory on automated emulation environments for grid interconnection testing at laboratory scale.
2021–2022
Comprehensive Privacy Framework for Streaming Data and Edge Computing $180,000
Principal Investigator — ORNL SEED
Outcome: Unified privacy framework combining differential privacy and secure computation for real-time data streams at the edge.
Principal Investigator — ORNL SEED
Outcome: Unified privacy framework combining differential privacy and secure computation for real-time data streams at the edge.
2021
Flexible Privacy-enabled Platform for AI Sensitive Applications $1,000,000
Task Lead — ORNL LDRD
Outcome: Modular platform enabling privacy-preserving AI pipelines for sensitive data domains including healthcare and national security.
Task Lead — ORNL LDRD
Outcome: Modular platform enabling privacy-preserving AI pipelines for sensitive data domains including healthcare and national security.
2021
A Semantic Graph Approach to Monitoring the Biological Literature for Biosecurity Applications $1,000,000
Task Lead — ORNL LDRD
Outcome: Knowledge-graph system for automated biosecurity surveillance, extracting and linking signals from the scientific literature at scale.
Task Lead — ORNL LDRD
Outcome: Knowledge-graph system for automated biosecurity surveillance, extracting and linking signals from the scientific literature at scale.
2020–2022
Privacy Challenges in Low-Shot Learning for GeoAI Applications $900,000
Co-Principal Investigator — ORNL LDRD
Outcome: Analysis of privacy risks in low-shot geospatial AI and mitigation strategies for sensitive geographic data use cases.
Co-Principal Investigator — ORNL LDRD
Outcome: Analysis of privacy risks in low-shot geospatial AI and mitigation strategies for sensitive geographic data use cases.
Industry
Real-Time and Physics Based Data Analytics for Thermal Power Plants $90,000
ML Advisor — Strategic Power Systems
Outcome: Physics-based ML analytics pipeline for real-time performance monitoring and anomaly detection in thermal power generation.
ML Advisor — Strategic Power Systems
Outcome: Physics-based ML analytics pipeline for real-time performance monitoring and anomaly detection in thermal power generation.
White Papers & Policy Responses
Invited contributions to agency solicitations, standards bodies, and policy processes.
2026
Request for Information Regarding Security Considerations for Artificial Intelligence Agents
Co-author — National Institute of Standards and Technology (NIST)
Arunachalam, H.B., Baburajan, B., Kotevska, O., et al. Input on security frameworks and threat models for agentic AI systems.
Co-author — National Institute of Standards and Technology (NIST)
Arunachalam, H.B., Baburajan, B., Kotevska, O., et al. Input on security frameworks and threat models for agentic AI systems.
2025
Privacy by Design in Distributed Edge Systems: Innovative Secure Workflows for Smart Cities
Author — IEEE Smart Cities Newsletter
Kotevska, O. Invited perspective on embedding privacy guarantees into edge computing architectures for smart city deployments.
Author — IEEE Smart Cities Newsletter
Kotevska, O. Invited perspective on embedding privacy guarantees into edge computing architectures for smart city deployments.
2024
Response to Notice of RFI Related to DOE’s Responsibilities on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence
Co-author — U.S. Department of Energy
Balaprakash, P., et al. Technical input on safety, security, and trustworthiness requirements for AI in DOE scientific mission contexts.
Co-author — U.S. Department of Energy
Balaprakash, P., et al. Technical input on safety, security, and trustworthiness requirements for AI in DOE scientific mission contexts.
2022
Comments on Advancing Privacy-Enhancing Technologies
Author — Office of Science and Technology Policy (OSTP), White House
Kotevska, O. Solo-authored response to the White House OSTP RFI on PETs, providing recommendations on differential privacy and federated learning standards for federal deployment.
Author — Office of Science and Technology Policy (OSTP), White House
Kotevska, O. Solo-authored response to the White House OSTP RFI on PETs, providing recommendations on differential privacy and federated learning standards for federal deployment.
2021
Challenges with Sensitive Data in Distributed Graph
Co-author — DOE ASCR Workshop on Cybersecurity and Privacy for Scientific Computing Ecosystems
Kotevska, O., Stanley, C., Michael, R., Kay, B., Sarwate, A., Kannan, R., & Tourassi, G. Privacy challenges and research directions for distributed graph analytics on sensitive scientific data.
Co-author — DOE ASCR Workshop on Cybersecurity and Privacy for Scientific Computing Ecosystems
Kotevska, O., Stanley, C., Michael, R., Kay, B., Sarwate, A., Kannan, R., & Tourassi, G. Privacy challenges and research directions for distributed graph analytics on sensitive scientific data.
2021
Addressing the Limitations to Distributed Learning Containing Sensitive Data
Co-author — DOE ASCR Workshop on Cybersecurity and Privacy for Scientific Computing Ecosystems
Michael, J.R., Stanley, C., Adamson, R., & Kotevska, O. Analysis of fundamental constraints in privacy-preserving distributed learning for scientific applications.
Co-author — DOE ASCR Workshop on Cybersecurity and Privacy for Scientific Computing Ecosystems
Michael, J.R., Stanley, C., Adamson, R., & Kotevska, O. Analysis of fundamental constraints in privacy-preserving distributed learning for scientific applications.