Headline

FSE/ FDE building AI agents & infra: VoiceScope (hallucination detection for voice AI), MCP Foundry (auto-generates & self-verifies MCP servers) |agent orchestration, RAG, multi-LLM systems |Open to FDE, Full Stack role

About

Know me:shivamchavan.is-a.dev Full Stack Engineer based in Bengaluru, building AI infra. Open to FDE, SDE, and AI/full-stack roles, on-site, hybrid, or remote. Best fit for: Forward Deployed Engineer, Full Stack, or Backend roles at AI-native or YC-backed startups. Hands-on across LLM providers, agent pipelines, and RAG, not just prompting. Location: Bengaluru, India · Available: Immediately · Open to: Relocation, remote, hybrid • VoiceScope: hallucination-detection platform for voice AI calls. 3-agent pipeline (transcribe, analyze, report) validated by a deterministic 7-layer harness instead of a second LLM grading the first one's output. 6 LLM providers (OpenAI, Anthropic, Gemini, Groq, Ollama, Mistral), 3 STT providers, RAG-grounded policy checks via ChromaDB, SSRF-protected webhook ingestion for 6 voice AI platforms. Deployed on Railway with Postgres. github.com/shivamchavan01/voiceScope • Consolidated Multi-Gym Dashboard (EazyTag): full-stack feature built solo. Angular PrimeNG frontend, Micronaut 4 backend with batched MongoDB queries across 50+ gym locations, a 3-factor health scoring engine, and automated anomaly alerts. • Event-driven notification pipeline at EazyTag: retry logic and failure isolation, zero message loss across 10+ tenants under degraded network conditions. • Currently building an AI agent project for OpenAI Build Week and the OpenAI x NamasteDev Codex Hackathon (July 2026), applying the same agent-pipeline and provider-abstraction patterns from VoiceScope. Before EazyTag, 11 months contributing to Eclipse JKube at the Eclipse Foundation. Refactored JUnit test suites across 15+ modules in a 50k+ line enterprise Java codebase, multiple PRs merged into production. Recognised Hacktoberfest 2024 Level 4 contributor. Stack: Java 17 · Spring Boot · Micronaut · Angular · TypeScript · Node.js · MongoDB · Kubernetes · Docker · AWS · AI Agents · LLM/RAG pipelines · ChromaDB I use Claude Code, GitHub Copilot, and Cursor daily to move faster without cutting corners on code quality. GitHub: github.com/shivamchavan01 Currently exploring full-time FDE, SDE, and AI/full-stack opportunities across Bengaluru, Pune, Hyderabad, and remote.

Location

India

AI stack

claude code

codex

grok

mcp

openai

opencode

Work experience

Full Stack Engineer at EazyTag, Oct 2025 - Present · 11 mo, Bengaluru, Karnataka, India, • Built the Consolidated Multi-Gym Dashboard end-to-end: Angular + PrimeNG frontend (health score rings, revenue trend charts, drill-down per gym) backed by Micronaut 4 APIs with batched MongoDB queries across 50+ locations, a nightly GymSnapshotScheduler, 3-factor health scoring engine, and a GymAlert system for automated anomaly detection. • Fixed a production PWA bug where users were stuck on stale app versions post-deploy — diagnosed root cause (SwUpdate never wired, CDN caching SW manifests for 1 hour), built PwaUpdateService with SW version detection and auto-reload, patched CDN cache headers in the deploy pipeline. Shipped as PR #642. • Architected an event-driven notification pipeline with retry logic and failure isolation — zero message loss across 10+ tenants under degraded network conditions. • Shipped 8+ REST APIs with full integration and regression test coverage using JUnit + TDD; reduced post-release incident rate by 40%, maintained 99%+ microservice uptime. • Collaborated with product managers and designers in Agile sprints end-to-end from requirement to production release; led peer and design code reviews.

Open Source Contributor at Hacktoberfest, Oct 2024 - Nov 2024 · 2 mos, Remote, Contributing to the JKube project as part of Hacktoberfest 2024. Worked on tasks like replacing deprecated methods and improving test cases. Submitted meaningful pull requests, enhancing project functionality. Collaborating with a global community of developers to improve open-source software. Leveraged Java and other tools to solve real-world coding challenges. Strengthened understanding of Git & GitHub workflows and version control practices.

Open Source Contributor | Eclipse JKube Project at Eclipse Foundation, May 2024 - Mar 2025 · 11 mos, Bengaluru, Karnataka, India, • Contributed to Eclipse JKube — a Maven/Gradle plugin automating Docker image builds and Kubernetes deployments for enterprise Java teams globally; worked in a 50k+ line multi-module codebase alongside Principal Engineers. • Refactored JUnit integration test suites across 15+ modules applying TDD and builder patterns — cut flaky CI pipeline failures by 60%, multiple PRs merged into production releases. • Recognised Hacktoberfest 2024 Level 4 contributor; collaborated through formal PR and design review processes in a globally distributed open source team.

Software Engineer Intern at KodNest, Jul 2023 - Oct 2024 · 1 yr 4 mos, Bengaluru, Karnataka, India, During my internship at KodNest Technologies, I actively acquired practical skills in Java, FullStack, and SQL, contributing significantly to real-world projects. I engaged directly with industry professionals, gaining valuable insights into the Software Development Life Cycle (SDLC) and honing expertise in Software Testing. I conducted manual testing on a simulated E-commerce website, leveraging tools like Jira to enhance quality assurance. This experience not only expanded my technical proficiency but also provided firsthand exposure to collaborative software development practices within a professional setting.

Education

Bachelor of Engineering, Electrical, Electronics and Communications Engineering, Bapuji Institute of Engineering & Technology, Davanagere

Diploma of Education, Electronic and communication Engineering, KLE polytechnic mahalingpur

Public links

github: https://github.com/shivamchavan01

shivamchavan.is-a.dev: https://shivamchavan.is-a.dev/