sukruthreddy Sukruthr

AI/ML Engineer focused on reliable LLM products, agent orchestration, and production deployment.

What I Build

I build LLM applications and agentic systems with clear APIs, practical workflows, and production-minded structure.
I focus on evaluation and reliability, especially hallucination benchmarking, output quality checks, and guardrail-minded design.
I ship across healthcare AI, retrieval systems, and cloud-native AI infrastructure from prototype through deployable services.

Project	What It Does	Value Signal
local-llm	FastAPI gateway in front of Ollama with API-key auth, rate limiting, and tunnel-ready access.	Turns local models into a deployable service boundary.
llm-hallucination-benchmark	Benchmarks hallucination behavior across 10 open models and 400 factual Q&A prompts in 5 domains.	Provides measurable reliability signals for model selection.
clinisight-ai	Medical diagnosis assistant API + MCP tool that extracts symptoms and retrieves supporting PubMed context.	Demonstrates healthcare-oriented AI with evidence retrieval.
agentic-travel-orchestrator	Multi-agent travel planning workflow using ADK-style orchestration and tool-driven execution.	Shows practical multi-agent orchestration patterns.
aws-bedrock-project	End-to-end AWS Bedrock RAG chatbot with Aurora pgvector, S3 knowledge base, and Terraform IaC.	Connects LLM capabilities to enterprise-style cloud architecture.
biomedical-knowledge-graph	Biomedical knowledge graph work across disease, gene, pathway, and drug relationships.	Applies graph-driven reasoning to domain-heavy biomedical data.

Also built: weather-mcp, adk-weather-agent, job-recommender, bAIWatch-tamuccHackathon

Agentic AI systems: robust orchestration, tool-use routing, and evaluation-aware agent workflows.
LLM reliability: benchmark design, failure analysis, and quality gates before deployment.
Production AI infrastructure: FastAPI services, local/private model gateways, and cloud deployment patterns.

Languages

AI / LLM

Infra

February 2026: refreshed profile repository and continued iteration on local-llm; active work in bAIWatch-tamuccHackathon.
January 2026: built and updated agentic-travel-orchestrator, adk-weather-agent, and clinisight-ai agent/API projects.
December 2025: shipped weather-mcp, llm-hallucination-benchmark, and aws-bedrock-project with evaluation and deployment-focused workflows.

Open to AI/ML Engineer opportunities focused on LLM systems, applied AI products, and production deployment.
Interested in collaborations around LLM evaluation, agentic architectures (MCP/ADK), and healthcare/knowledge-graph AI.