Sitemap
A list of all the posts and pages found on the site. For you robots out there, there is an XML version available for digesting as well.
Pages
Posts
portfolio
publications
Target-Agnostic Adversarial Attacks on Language Understanding Models
Published in arXiv preprint (2106.07047), 2021
Proposes adversarial attacks on language-understanding models that do not assume access to the victim’s label space or target outputs. The attacks transfer across tasks and demonstrate systemic robustness gaps in standard NLU model families.
Recommended citation: Chauhan, J., Bhukar, K., Kaul, M. (2021). "Target-Agnostic Adversarial Attacks on Language Understanding Models." arXiv preprint 2106.07047.
Download Paper
End-to-End Deep Reinforcement Learning for Conversation Disentanglement
Published in Proceedings of the 37th AAAI Conference on Artificial Intelligence (AAAI 2023), 2023
Formulates conversation disentanglement as a sequential decision-making problem and trains an end-to-end deep RL agent to cluster interleaved multi-party chat messages into coherent threads.
Recommended citation: Bhukar, K., Kumar, H., Raghu, D., Gupta, A. (2023). "End-to-End Deep Reinforcement Learning for Conversation Disentanglement." Proceedings of the 37th AAAI Conference on Artificial Intelligence.
Download Paper
Systems and Methods to Automatically Generate Close Notes from SRE Interactions for Downstream AIOps Tasks
Published in U.S. Patent Application, filed 2023 — Application No. US20250182053A1, 2023
U.S. patent application (2023). Automatically synthesizes structured close notes from free-form SRE interaction transcripts, producing reusable artifacts that power downstream AIOps tasks such as retrieval, classification, and recommendation.
Recommended citation: Bhukar, K., Kumar, H., Azad, A., Joshi, S. "Systems and Methods to Automatically Generate Close Notes from SRE Interactions for Downstream AIOps Tasks." U.S. Patent Application US20250182053A1, filed 2023.
Download Paper
Systems and Methods to Recommend or Revoke Agents Guided by the AIOps Conversation Context
Published in U.S. / CN / JP Patent Applications — Application No. US20250117282A1, 2023
U.S. patent 2023, China/Japan 2024. Dynamically recommends or revokes AIOps agents based on the evolving context of an operator-agent conversation.
Recommended citation: Bhukar, K., Kumar, H., Azad, A., Joshi, S. "Systems and Methods to Recommend/Revoke Agents Guided by the AIOps Conversation Context." U.S. Patent Application US20250117282A1, filed 2023; CN/JP applications 2024.
Download Paper
Learning Representations on Logs for AIOps
Published in Proceedings of the 16th IEEE International Conference on Cloud Computing (IEEE CLOUD 2023), 2023
Pre-trains a log-domain language model on large-scale production log corpora and demonstrates transfer to downstream AIOps tasks such as anomaly detection and log parsing.
Recommended citation: Gupta, P., Kumar, H., Kar, D., Bhukar, K., Aggarwal, P., Mohapatra, P. (2023). "Learning Representations on Logs for AIOps." Proceedings of the 16th IEEE International Conference on Cloud Computing (CLOUD).
Download Paper
Dynamic-X-Y: A Tool for Learning Dynamic Alert Suppression Policies in AIOps
Published in Proceedings of the 16th International Conference on COMmunication Systems and NETworkS (COMSNETS 2024) — Workshop, 2024
Dynamic-X-Y is the production tooling companion to the dynamic alert-suppression research, packaging the policy-learning loop for deployment on live observability streams.
Recommended citation: Bhukar, K., Kumar, H., Arora, R., Mahindru, R., Nagar, S., Thornhill, M., Manning, I., Buggins, J., Cook, S., Paradkar, A., Aggarwal, P. (2024). "Dynamic-X-Y: A Tool for Learning Dynamic Alert Suppression Policies in AIOps." Proceedings of the 16th International Conference on COMmunication Systems and NETworkS (COMSNETS).
Download Paper
Dynamic Alert Suppression Policy for Noise Reduction in AIOps
Published in Proceedings of the 46th International Conference on Software Engineering: Companion Proceedings (ICSE-Companion 2024), 2024
Learns dynamic, per-signal alert-suppression policies to reduce noise in AIOps pipelines. Adapts suppression decisions to evolving signal characteristics, cutting duplicate alerts while preserving fidelity on actionable incidents.
Recommended citation: Bhukar, K., Kumar, H., Mahindru, R., Nagar, S., Aggarwal, P., Arora, R., Paradkar, A. (2024). "Dynamic Alert Suppression Policy for Noise Reduction in AIOps." Proceedings of the 46th International Conference on Software Engineering: Companion Proceedings (ICSE-Companion). DOI: 10.1145/3639477.3639752.
Download Paper
Systems and Methods for Determining Log Templates of a Computing Device for Log Templatization
Published in U.S. Patent Application, filed 2024 — Application No. US20260030134A1, 2024
U.S. patent application (2024) — Application No. US20260030134A1. Determines log templates from computing-device logs to drive efficient and accurate log templatization for downstream analytics.
Recommended citation: Bhukar, K., Gupta, P., Kumar, H., Verma, M. "Systems and Methods for Determining Log Templates of a Computing Device for Log Templatization." U.S. Patent Application US20260030134A1, filed 2024.
Download Paper
Systems and Methods for Dynamic Policy Management for Events and Alerts Noise Reduction in IT Systems
Published in U.S. Patent No. US12277025B2 (granted 2025), 2025
U.S. Patent No. US12277025B2 (granted 2025). Learns dynamic per-signal alert-suppression policies from historical observability streams and pushes them as runtime policies.
Recommended citation: Bhukar, K., Nagar, S., Kumar, H., Aggarwal, P., Mahindru, R., Paradkar, A., Manning, I., Thornhill, M., Arora, R., Hussey, S., Forti, F. "Systems and Methods for Dynamic Policy Management for Events and Alerts Noise Reduction in IT Systems." U.S. Patent No. US12277025B2, granted 2025.
Download Paper
LogAn: An LLM-Based Log Analytics Tool with Causal Inferencing
Published in Proceedings of the 25th ACM/SPEC International Conference on Performance Engineering (ICPE 2025), 2025
LogAn is an LLM-driven log analytics tool that uses causal inference to explain production incidents from raw log streams. It combines structured log representations with causal reasoning to surface likely root causes and accelerate triage in large-scale cloud systems.
Recommended citation: Gupta, P., Bhukar, K., Kumar, H., Nagar, S., Mohapatra, P., Kar, D. (2025). "LogAn: An LLM-Based Log Analytics Tool with Causal Inferencing." Proceedings of the 25th ACM/SPEC International Conference on Performance Engineering (ICPE). DOI: 10.1145/3680256.3721246.
Download Paper
Systems and Methods for Logline Enrichment for Error Analysis
Published in U.S. Patent Application, filed 2025, 2025
U.S. patent application (2025). Enriches raw log lines with semantic and contextual signals to improve downstream error analysis and root-cause identification.
Recommended citation: Bhukar, K., Kumar, H., Nagar, S., Gupta, P., Samanta, S., Mohapatra, P., Kar, D., Aggarwal, P. "Systems and Methods for Logline Enrichment for Error Analysis." U.S. Patent Application, filed 2025.
Systems and Methods for Explainability-Driven Control of an Alert Suppression Policy
Published in U.S. Patent No. US12536061B1 (granted 2026), 2026
U.S. Patent No. US12536061B1 (granted 2026). Uses explainability signals to guide runtime control of an alert-suppression policy, tying each suppression decision to the evidence that supports it.
Recommended citation: Bhukar, K., Kumar, H., Nagar, S., Mahindru, R., Paradkar, A., Manning, I., Hussey, S. "Systems and Methods for Explainability-Driven Control of an Alert Suppression Policy." U.S. Patent No. US12536061B1, granted 2026.
Download Paper
Scalable and Efficient Large-Scale Log Analysis with LLMs: An IT Software Support Case Study
Published in Proceedings of the 40th AAAI Conference on Artificial Intelligence (AAAI 2026), 2026
Shows how LLM-based log analytics can be scaled to IT software-support workloads. Presents an end-to-end case study on millions of log lines across hundreds of enterprise incidents, reporting engineer-time savings and deployment learnings.
Recommended citation: Gupta, P., Bhukar, K., Kumar, H., Nagar, S., Mohapatra, P., Kar, D. (2026). "Scalable and Efficient Large-Scale Log Analysis with LLMs: An IT Software Support Case Study." Proceedings of the 40th AAAI Conference on Artificial Intelligence.
Download Paper
talks
End-to-End Deep Reinforcement Learning for Conversation Disentanglement
Published:
Presented the AAAI 2023 paper on end-to-end RL for conversation disentanglement at the AAAI-23 main conference in Washington, D.C.
IBM Research Industry Expo — NeurIPS 2023
Published:
Represented IBM Research at the NeurIPS 2023 industry expo in New Orleans.
Dynamic Alert Suppression Policy for Noise Reduction in AIOps
Published:
Presented at the ICSE 2024 companion track in Lisbon — the alert-suppression work done at IBM Research.
teaching
Artificial Intelligence — Teaching Assistant
Undergraduate course, Department of Computer Science and Engineering, IIT Hyderabad, 2019
Teaching Assistant for Artificial Intelligence, Spring 2019, under Prof. P. K. Srijith. Topics included classical search, adversarial search, logic and reasoning, planning, and an introduction to probabilistic reasoning.
Data Structures — Teaching Assistant
Undergraduate course, Department of Computer Science and Engineering, IIT Hyderabad, 2020
Teaching Assistant for Data Structures, Fall 2020, under Prof. Karteek Sreenivasaiah. Covered arrays, lists, trees, heaps, graphs, and hashing; ran weekly doubt-clearing sessions and graded programming assignments in C++.
Foundations of Machine Learning — Teaching Assistant
Graduate course, Department of Computer Science and Engineering, IIT Hyderabad, 2021
Teaching Assistant for Foundations of Machine Learning, Fall 2021, under Prof. Vineeth N. Balasubramanian. Helped design and grade assignments and led tutorial sessions on probabilistic ML, linear models, kernels, and deep learning foundations.