Senior software engineer and systems researcher working on reliable cloud infrastructure, identity and agent systems, distributed inference, and adaptive cloud systems.
- Distributed LLM inference and request scheduling
- Reliable LLM agent systems and structured tool-output validation
- Kubernetes and KEDA autoscaling
- Heterogeneous event-driven workloads
- Cloud reliability, observability, and performance evaluation
-
Executable Task-Equivalence Checks for Structured LLM Tool Outputs
Reproducibility package for executable task-equivalence checks, external-transform validation, benchmark data, and negative-result analysis. -
Semantic Cost-Aware Autoscaling Prototype
KEDA reference implementation for evaluating workload-aware scaling against queue-length baselines.
Both repositories include citation metadata and reproducibility checks.
- Invariant-Preserving Compression of Structured Tool Outputs for LLM Agents. Manuscript submitted to The Journal of Systems & Software.
- Safety-Floored Work-Aware Autoscaling for Heterogeneous Event-Driven Services. Manuscript submitted to Software: Practice and Experience.
- Peer-review service for Web of Science-indexed journals.
- NSF reviewer profile.
- Program-committee service and nominations in distributed systems, cloud computing, software architecture, reliability, DevOps, and observability.