Kyle R Krcmaric

Software Engineer

Atlanta, GA, USA
kyle@krcmaric.com

GitHub

Backend-focused software engineer with 6+ years of experience building production Ruby on Rails APIs, Python ML pipelines, and AWS cloud infrastructure. Specialized in multi-agent LLM systems with GraphRAG, containerized microservices on ECS Fargate, and high-scale backend services for telecommunications platforms. Deep expertise in infrastructure-as-code (AWS CDK, CloudFormation), graph databases (Neo4j), and distributed systems. Passionate about backend architecture, platform engineering, and bridging production systems with cutting-edge AI capabilities through robust API design and scalable infrastructure.

Experience

, Blacksburg VA (Remote) - Software Engineer III

February 2022 - Present

GraphRAG Initiative

Secured C-level buy-in from CEO and VP of Technology for GraphRAG knowledge graph initiative, moving from experimental R&D to productization with Q1 2026 MVP target
Architected and deployed end-to-end containerized stack on AWS ECS Fargate using CDK infrastructure-as-code, including web server, custom vector search layer (ChromaDB), Neo4j graph database, and full networking/DNS configuration
Pioneered graph database adoption to modernize legacy data model, demonstrating superior query capabilities with Cypher over traditional relational approaches for company's data structure

V2 API Leadership

Led technical architecture and optimization for V2 API launch serving 61K requests/day, coordinating 3-4 engineers across agent and self-serve teams
Reduced average API response time by 50% through request profiling with Vernier and query optimization, enabling customers to accomplish in one V2 query what previously required multiple V1 queries
Drove API adoption with customer integrations gaining momentum over past 1-2 quarters

Infrastructure & Messaging

Architected Twilio SMS migration for 80K messages/week from AWS, owning ground-up integration with Twilio API, callback status handling, and feature-flagged gradual rollout strategy
Led regulatory compliance improvements and short code acquisition, documenting messaging practices and driving organizational buy-in for recommendations, reducing spam filtering risk and enabling higher throughput
Maintain production systems at scale: 65K DAU, 30 req/sec average, processing 1.5M Sidekiq jobs/day with 99.998% success rate

Platform & Developer Experience

Established containerization standards by creating VS Code devcontainer configurations adopted by new engineers, with comprehensive README documentation lowering barrier to entry for backend infrastructure
Perform cross-repository code reviews for Rails API, Gateway, and serverless Lambda/CloudFormation codebases

, Blacksburg VA (Remote) - Innovation Engineer

March 2020 - February 2022

Launched semantic search to production enabling call center agents, consumers, and V2 API to discover self-support content, following years of R&D groundwork.
Developed custom NER model with annotated dataset for device identification, solving previously impossible challenge of confidently parsing user device references across search queries.
Prototyped solutions supporting market expansion beyond agent and self-serve channels into deep API integration opportunities.

, Blacksburg VA (Remote) - Associate Innovation Engineer

January 2019 - March 2020

Collaboratively created and rapidly iterated on research & development style prototypes to support long term company roadmap goals and market expansion efforts.

Licenses & certifications

Neo4j Certified Professional

Neo4j

Issued October 2019

Credential ID 16899288

Skills

AWS Cloud Infrastructure (CDK, Lambda, ECS Fargate, CloudFormation)

Infrastructure-as-code using CDK and CloudFormation. Experience with ECS Fargate, Lambda, EC2, VPC, Load Balancers, RDS, DynamoDB, IAM, and Secrets Manager.

Docker & Containerization

Multi-stage Dockerfiles for Ruby, Python, TypeScript, and Node.js. ECS Fargate orchestration, Docker Compose for local development, and CI/CD pipeline automation.

Ruby on Rails

Primary backend language. RESTful APIs, Sidekiq background jobs, Elasticsearch, ActiveRecord, Passenger/Nginx deployment, and Rubocop.

Python

NLP, machine learning, REST APIs, and AWS Lambda for ML pipelines and microservices.

TypeScript

React frontends, AWS CDK infrastructure-as-code, multi-agent AI orchestration, and type-safe APIs.

Neo4j & GraphRAG

Neo4j Certified Professional. Graph database design, Cypher optimization, knowledge graphs, vector indexes, and LLM integration.

LLM Agent Development

Multi-agent systems with Google ADK and Model Context Protocol. Claude, GPT, and Gemini API integration.

Networking / Web Hosting

Self-hosted and cloud hosted websites with DNS configuration (you're here now!).

Notable Projects

GraphRAG Multi-Agent AI System | AWS ECS Fargate, Neo4j, ChromaDB, Google ADK, MCP

Heavily involved in experimental R&D initiative securing CEO-level buy-in for productization. Built multi-service containerized architecture deployed via AWS CDK. Moving to production team to lead MVP delivery.

V2 API Performance & Architecture | Ruby on Rails, Vernier, MySQL

Technical lead for cross-team API redesign serving 61K requests/day. Reduced response times 50% through query optimization. Simplified customer integration from multiple V1 queries to single V2 calls.

Twilio SMS Migration & Compliance | Twilio API, Ruby, Feature Flags

Sole engineer for initial migration from AWS to Twilio. Architected callback handling, status integration, and feature-flagged rollout. Led compliance documentation driving short code acquisition for improved deliverability.