I build production AI agents and autonomous engineering platforms — RAG systems, MCP integrations, and Bedrock-powered workflows that take tickets to production in minutes. 15+ years of experience across cloud infrastructure, security, and AI/ML, with a focus on making AI genuinely useful in the workflows that developers and security engineers live in every day.

Single-agent, multi-agent, MCP-based, browser, voice, and local agents — including Bedrock Converse API agents, Step Functions multi-agent orchestration, and hybrid RAG+MCP architectures

Ticket-to-PR pipelines, Jira lifecycle automation (ticket → semantic search → code gen → PR → status updates), GuardDuty AI triage with auto-remediation, and end-to-end ticket-to-production workflows

Structuring agents, integrating tools, managing logs, evaluating, and planning — plus Bedrock IAM/inference profile management and prompt engineering for code generation

From simple chains to agentic RAG, hybrid search, and local retrieval setups — including Aurora pgvector code intelligence over GitHub orgs, Slack/Confluence helpbots with confidence-tiered responses, and RAG vs MCP architectural tradeoffs

LiteLLM proxy layers over Bedrock, REST API wrappers via Terraform, multi-model routing (Claude 4.x via inference profiles), and consolidated AWS billing for AI workloads

Integrations for GitHub, PDFs, videos, and research papers — extended with Confluence, Jira, Slack, and CloudTrail-as-corpus

GitHub, AWS, Slack, Confluence, JIRA — plus ServiceNow Discovery, Tenable, SentinelOne, Illumio CloudSecure, and IONIX

Agents for vulnerability triage, compliance checks, and threat intel summarization — including supply chain compromise audits (Trivy incident response), dangling Route53 NS detection, CloudTrail IAM anomaly detection via LLM, and SMS pumping/AIT defense

Infrastructure scanning, secrets detection, and auto-remediation workflows — plus PatchOps Command Center, ECR lifecycle automation, EC2 patch compliance, Identity Center recertification, and Tenable-to-EC2 auto-remediation

Cloud billing anomaly detection with multi-factor scoring, RI-aware false-positive filtering, alert fatigue mitigation, and ServiceNow cost/value analysis

Terraform modules (ALB, networking), AWS Control Tower environments, Account Factory for Terraform (AFT), StackSets, SCP debugging, and cross-account role patterns

TradingView Pine Script v6 intraday signal bots, multi-tier partial exit logic, signal scoring on 2m/10m charts, and US equities entry/exit refinement

Working with Llama, SLMs, and other open-source models

💡 Always exploring how AI can make engineering faster, secure, safer, and more efficient.