Aman Sriven

sriven.aman@gmail.com(214) 991-3721linkedin.com/in/aman-srivengithub.com/amansriven

Skills

PythonC++CJavaTypeScriptJavaScriptSQLGoSolidityHTML/CSSFastAPIReactNode.jsNext.jsPyTorchSparkDelta LakeMongoDBPostgreSQLKafkaAWSGCPAzureDatabricksKubernetesDockerHelmArgoCDCI/CDGitHub

Experience

Software Engineering Intern

Humana

2026-05 – 2026-08 · Louisville, KY

  • Architected a distributed AI gateway routing 450K+ requests/day on Azure Kubernetes via Envoy, Helm, ArgoCD
  • Secured internal APIs behind MCP servers with per-user tool scoping; tool-selection accuracy 68% to 89%
  • Rebuilt RAG ingestion, embedding, and ranking Spark pipelines on Databricks Delta Lake; retrieval recall up 40%

Artificial Intelligence Engineering Intern

Jaggaer

2025-06 – 2025-08 · Durham, NC

  • Engineered a Python retrieval and orchestration agent over a 4M+ supplier corpus; cut procurement cycle time 40%
  • Raised supplier-matching and classification accuracy 25% over rules-engine baselines on 10,000+ labeled examples
  • Automated model training, retraining, and inference pipelines over a vector search index in GitHub Actions CI

Undergraduate Researcher

Texas A&M University

2024-12 – 2025-05 · College Station, TX

  • Developed an ensemble PyTorch classifier detecting 11 security vulnerability classes in 100k+ smart contracts
  • Paired residual connections and multi-head attention with adaptive focal loss, reaching 94–96% balanced accuracy
  • Beat published benchmark F1 of 0.88 across all 11 vulnerability labels; won 1st of 40 at 2025 research symposium

Projects

◎

Delta Code

Full-stack Dependabot-style API checker diffing base and PR behavior to block breaking changes

  • Shipped a full-stack Dependabot-style API checker diffing base and PR behavior to block breaking changes
  • Integrated GitHub Apps with webhooks, token checkout, and Check Runs on a queued PostgreSQL backend worker
  • Layered LLM test-case generation and reviewer-facing explanations; cut median time-to-close 30% across 50+ PRs
◆

Odyssey AI

9-agent LLM pipeline triaging denied healthcare claims and auth gaps; won 1st at Humana hackathon

  • Delivered a 9-agent LLM pipeline triaging denied healthcare claims and auth gaps; won 1st at Humana hackathon
  • Demoted the solvability router from an LLM call to deterministic code so agents explain findings, not decide facts
  • Serialized six concurrent tool calls from two parallel agents, trading 5s of latency to clear Vertex AI rate limits
◈

Ahvaan

Customer-facing iOS and Android RSVP app with phone-number auth and calendar sync for 600+ users

  • Launched a customer-facing iOS and Android RSVP app with phone-number auth and calendar sync for 600+ users
  • Designed a secure Node.js REST backend routing invites to push, WhatsApp, or SMS without exposing contacts
◈

Ocean Robotics Submarine

Real-time OpenCV perception pipeline for an autonomous submarine

  • Fused camera, sonar, and IMU inputs into a real-time OpenCV perception pipeline for an autonomous submarine
  • Trained a PyTorch vision model for obstacle detection to 50 ft, feeding live signals into the submarine control loop

Education

Texas A&M University

Career highlights

May 2026

  • Started: Software Engineering Intern at Humana

    Humana

June 2025

  • Started: Artificial Intelligence Engineering Intern at Jaggaer

    Jaggaer

December 2024

  • Started: Undergraduate Researcher at Texas A&M University

    Texas A&M University

Built with Round Zero