Software
& AI
Engineer
3+ years building and operating high-throughput Java/Spring Boot microservices in regulated financial and enterprise environments (USAA, AT&T). Currently engineering production-quality open-source Python infrastructure at Arizona State University. MS in Data Analytics from ASU. Open to AI/LLM, Backend SWE, and Data roles.
Tarun Sai Marisetti
Backend Engineering
3+ years designing production-grade Java/Spring Boot microservices with fault-tolerance patterns, REST APIs, and event-driven Kafka pipelines for AT&T and USAA workloads.
LLM & AI Engineering
Built RAG pipelines with LangChain + ChromaDB, developed multi-agent systems using GPT-4o and Claude APIs, and leveraged generative AI as a daily productivity accelerator across the full development workflow.
Data Analytics
MS in Data Analytics from ASU. Built predictive ML models (LightGBM, SHAP) on 900K+ crash records, applied unsupervised clustering to NASA exoplanet datasets, and delivered interactive data storytelling with D3.js and Three.js. Skilled in end-to-end analytical pipelines from raw data to actionable insight.
- Engineering and packaging Aurora, a published atmospheric retrieval framework, into a production-quality open-source Python library with modular architecture and clean public APIs.
- Building pytest test suite with CI/CD integration via GitHub Actions; enforcing coverage thresholds and automated quality gates on every commit.
- Authored Read the Docs site (API reference, conceptual overview, installation guide) and Jupyter tutorial notebooks to lower onboarding friction for new research users.
- Collaborating with cross-functional research team to define project scope, milestones, and deliverables toward open-source release.
- Designed and operated real-time Order Service exposing developer-facing APIs managing order lifecycle across 800K–1M calls/month; implemented idempotent workflows preventing duplicate billing and payment invocations.
- Designed async processing pipeline using Spring thread pools and Java concurrency primitives to sustain production-level throughput, reducing p99 latency under peak load.
- Diagnosed and resolved production thread pool exhaustion causing latency spikes under peak load; restructured executor configuration and authored runbook adopted by on-call rotation, reducing MTTR.
- Designed fault-tolerant patterns using Resilience4j circuit breakers with exponential backoff retries, preventing cascading failures during partial outages.
- Participated in Strangler Fig migration from legacy Rails monolith to Spring Boot; owned resolution of Spring AOP proxy bypass causing silent Async/Transactional failures by restructuring bean dependencies, preventing silent data consistency issues.
- Designed scheduled orchestration jobs with @Transactional boundaries and idempotency guards managing enrollment state transitions across multiple tables, preventing duplicate processing on retry.
- Contributed to Spring Boot orchestrator service handling downstream API coordination and financial data processing under strict regulatory compliance, data access controls, and audit requirements in a zero-tolerance financial environment (USAA).
- Implemented access control enforcement and least-privilege principles across backend services; managed Tier-3 production incidents via ServiceNow in a zero-tolerance security environment.
- Wrote JUnit/Mockito tests at ~80% coverage; participated in Agile ceremonies, code reviews, and production support rotations.
Cloud-Native Microservices Reliability Platform
Production-grade simulation of distributed system control-plane behavior: service discovery, auth, routing, and incident automation. Implements idempotent APIs, circuit breakers, rate limiting, and Kafka event-driven pipelines.
FinSight — Autonomous Financial Due Diligence Agent
Built a 6-agent hierarchical AI system that fetches SEC 10-K/10-Q filings via EDGAR API, parses 300-page documents, and generates semantic chunks — reducing financial due diligence from 40+ hours to under 5 minutes. Implemented RAG with query rewriting, LLM reranking over ChromaDB, and a multi-model ensemble (GPT-4o + GPT-4o-mini) producing structured risk scores, confidence levels, and filing citations.
Exoplanet Habitability Atlas — ASU Research Collaborator
Interactive research dashboard over NASA's 4,500+ planet archive with research-grade metrics (TSM/ESM, Kopparapu HZ classification) used in JWST telescope time allocation. Engineered multi-tab D3.js workspace with sortable tables, mass-radius diagrams, and composite Golden Targets scoring (ESI × TSM). Integrated AI-generated science briefs via Claude and GPT-4 APIs. Built Python ML pipeline for mass imputation via Random Forest across ~68% of planets with missing measurements.
Pac-Man AI Agent — Search & Reinforcement Learning
Monte Carlo Tree Search agent for heuristic ranking and decision-making under uncertainty. Reward optimization analogous to search relevance tuning in recommendation systems.
Chicago Traffic Analytics & Predictive Modeling Platform
Processed 2.4M+ records (~900K crashes, 1.5M enforcement) from Chicago open datasets, integrated with U.S. Census demographics for enriched modeling. Trained LightGBM crash severity classifier optimized for PR-AUC and F2-score under class imbalance; applied SHAP for top risk factor analysis. Mapped model outputs to Chicago geographic zones for non-technical stakeholders.
Open to AI, Backend & Data Roles
Currently on F1 OPT, authorized to work in the US until Jan 2027 with STEM OPT extension available for +2 years. No employer sponsorship cost during OPT period. Based in Tempe, AZ, open to remote roles nationwide.