goutamadwant - Overview (original) (raw)

Hi, I'm Goutam

Principal Engineer building agent reliability tooling, AI infrastructure, and cloud-native distributed systems.

GitHub Developer Program member | AI agents | MCP security | LLM infrastructure | AWS | Java | Python

I work on open-source tools for testing, securing, and operating AI systems. My focus is practical research: methods that can be tested, reproduced, and used by engineers building real platforms.

Current Open Source Projects

Agent Testing & MCP Security

• 🧪 toolcallcheck - offline regression tests for tool-calling agents.
• 🔌 mcp-pool - async connection pooling for MCP client sessions.
• 🛡️ mcp-shield-pii - PII detection and reversible redaction for MCP traffic.
• 🚦 mcp-egress-guard - policy enforcement and audit logs for MCP egress.

LLM Infrastructure

• 🧭 kvfleet - KV-cache-aware routing for self-hosted and hybrid LLM fleets.
• 🧬 vectormigrate - safer embedding and vector-index migrations.
• 🧩 promptspecj - OpenAPI-style prompt contracts for Java and Spring AI.

Developer & Data Tools

• 📈 bigocheck - zero-dependency empirical complexity checks.
• ⚡ pdperf - static lints for slow Pandas anti-patterns.
• 🏗️ pydanticforge - infer Pydantic v2 models from evolving JSON streams.
• ✨ schemaglow - schema diff and contract-drift detection.
• 🗺️ tracemap - terminal and web visualization for traceroute data.

Research Paper Implementations

• 📚 rag-evidence-coverage-evaluator - RAG grounding and evidence-coverage evaluation.
• 🔗 rag-chain-of-logic-coverage - reasoning-validity checks for retrieval workflows.
• 🧬 orthogonal-low-rank-editing - orthogonal low-rank updates for neural knowledge editing.
• 🧠 amorphous-intelligence-experiments - local-LLM experiments on drift, divergence, and semantic skew.

Research & Engineering Focus

• Agent reliability: deterministic tool-call testing, privacy, and egress policy.
• AI infrastructure: LLM routing, MCP session management, and vector migration.
• Developer tooling: prompt contracts, complexity checks, static analysis, and schema drift.
• Cloud-native systems: reliable services, event-driven platforms, and operational automation.

Collaboration

I am interested in focused open-source work around agent security, AI infrastructure, developer productivity, and cloud-native AI platforms.

GitHub: github.com/goutamadwant