Research accepted — IJCAI-ECAI 2026, Demo Track

Kavya Gautam

Computer Science @ Mississippi State University. Building distributed systems, AI pipelines, and research tools. Open to Summer / Fall 2026 internships.

Research & Work

Undergraduate Research Assistant
AI Research (AIR) Lab — Mississippi State University
Feb 2026 – Present
Project: LLM-to-Map — EV Grid Simulation Platform
  • Research accepted for demonstration at IJCAI-ECAI 2026 (Demo Track)
  • Migrated architecture from HTTP polling to fully event-driven WebSocket system
  • Integrated agentic AI chatbot with 26 OpenAI function-calling tools for live simulation control
  • Scaled frontend to 700+ concurrent vehicles with 3D terrain rendering; containerized with Docker Compose
WebSocketsOpenAI APIDockerFastAPI
Undergraduate Research Assistant
Office of Research & Economic Development — Mississippi State University
Jun 2025 – Present
Project: NSF SECURE Research Intelligence Platform
  • Lead developer integrating 8+ datasets (ROR, IPEDS, NSF HERD, OpenAlex, Crossref, USASpending)
  • Built enrichment pipelines for 1,400+ publications with author affiliations and funding attribution
  • Full-stack webapp with choropleth maps and topic-hierarchy network visualizations
TypeScriptReactPostgreSQLREST APIs
AI Fellow
Mississippi AI Innovation Hub
Dec 2025 – May 2026
Project: Legal Compliance RAG System
  • Scaled RAG pipeline from 1-state prototype to 7 states and 5,000+ legal documents
  • Built web crawlers using Playwright and Scrapy; enforced hallucination guardrails via citation-only generation
  • Configured AWS Bedrock with OpenSearch Serverless and hierarchical chunking
AWS BedrockOpenSearchPlaywrightScrapyStreamlit

Selected Work

Distributed Remote Code Execution Engine

↗ github

Fault-tolerant system for sandboxed execution of untrusted code. Reduced median latency 99% (2.6s → 24ms) via async I/O and connection pooling. Secured with Rootless Podman, network air-gapping, and custom Seccomp profiles.

PythonFastAPIWebSocketsRabbitMQPodman

Vision Transformer for Audio Classification

↗ github

Adapted a pre-trained ViT for audio spectrogram classification on UrbanSound8K, achieving 71.2% accuracy without audio-specific pretraining. Trained on NVIDIA A100 GPUs via Slurm on MSU's HPC cluster.

PyTorchViTSlurmHPCNVIDIA A100

Technical Stack

Languages
PythonC++SQLTypeScriptJavaScriptBash
Cloud & DevOps
AWS (S3, EC2, Bedrock)DockerPodmanKubernetesGitHub Actions
Backend & Distributed
FastAPIRabbitMQPostgreSQLWebSocketsMicroservices
ML & Data
PyTorchNumPyPandasScikit-learnOpenSearchRAG
Observability
PrometheusGrafanaLocust

Community

President
ACM SIGHPC — Mississippi State University
Sep 2024 – May 2026
  • Led MSU's HPC community as chapter President
  • Organized biweekly workshops on parallel computing and HPC tooling
  • Previously served as VP managing chapter logistics and communications