{ }
AI Platforms
AWS Bedrock AgentCoreStrands AgentsFastAPI MCPLLM workflowsAgent runtime patternsPrompt systems
Available for the right opportunity
howdy, i'm
Rudy Pena
Staff Software Engineer focused on applied AI systems and ML platform infrastructure. At DTN, I build secure workflows with AWS Bedrock AgentCore, Strands agents, and FastAPI MCP services. I use Python and Node.js CLI/SDKs to help engineers move from prototype to customer-ready agents faster.
interactive terminal - try it
rudy@portfolio ~ % type "help" to explore
About
Building systems developers trust.
I am a Staff Software Engineer focused on applied AI platforms, ML infrastructure, and the developer experience around them. Most of my recent work sits at the intersection of AWS Bedrock AgentCore, Strands agents, FastAPI MCP services, LLM workflows, and the Python/Node.js SDKs that help teams turn prototypes into useful products.
Before the AI platform work, I spent years shipping developer-facing products, component libraries, web applications, and infrastructure at DTN, Anaconda, IBM, and Bitfusion.
The common thread is pretty simple: I like making complicated work feel approachable for the people who have to build on top of it.
10+Years building software
LLMSystems and workflows
MCPTooling and integrations
PythonAI tooling
Currently at DTN
Designing AI platform infrastructure, AWS Bedrock AgentCore workflows, Strands agents, FastAPI MCP services, and self-service paths for product teams.
Skills
Tools I use to build.
∑
ML & Retrieval
PythonFastAPIPyTorchTransformersEmbeddingsVector searchEvals
[]
Cloud Systems
API Gateway v1/v2ECSLambdaS3ECRAurora/PostgresOpenSearchTerraform
<>
Product Engineering
Next.jsReactTypeScriptNode.js APIsAuth0Design systemsTailwind CSS
$_
Developer Tooling
Python SDKsPython CLIsNode.js toolingRoles/permissionsCI/CDGitLab pipelines
Experience
Roles that shaped the work.
Staff Software Engineer, AI Platform
DTN
- Designed a secure, identity-aware AI platform on AWS that enables teams to prototype and deploy agent-driven workflows with shared and custom MCP tooling.
- Built runtime and deployment patterns for AWS Bedrock AgentCore, Strands-based agents, and FastAPI MCP services using API Gateway, WAF, Lambda authorizers, ECR, S3, Aurora/Postgres, and OpenSearch.
- Created Python CLI/SDK tooling to bootstrap agent services, standardize configuration, and reduce time-to-prototype for customer-facing AI use cases.
- Delivered product and platform surfaces in Next.js for agent creation, configuration, and interaction across engineering and product teams.
- Delivered production AI features with AWS Bedrock Claude and Titan for summarization, analysis, and insight generation in DTN's Grain Intelligence platform.
AWS Bedrock AgentCoreFastAPI MCPPython SDKsNext.jsIdentity/AuthOpenSearch
Software Engineer III
Anaconda
- Built and shipped core Anaconda.com features with Angular and FastAPI for a developer platform used by millions.
- Led the Karma-to-Jest migration and helped raise test coverage from 2% to 60%.
- Drove MVP initiatives that grew into a meaningful share of the product portfolio, including partner work with HP, Snowflake, and Microsoft.
AngularFastAPIJestPythonProduct MVPs
Senior UI Developer
IBM
- Developed a large portion of IBM's internal component library in Angular.
- Coordinated work across Austin and India teams to break down epics and deliver MVP scope.
AngularComponent librariesDesign systems
Software Engineer
Bitfusion
- Built Go CLI tooling for distributed machine learning workloads with Cobra and Viper.
- Developed full-stack features with Go, Node.js, and React for managing ML jobs and clusters.
GoNode.jsReactCLI toolingML workloads
Projects
Proof through working projects.
Wavves - Music Lab
A music AI lab for concert prep, smart playlists, and song discovery. It combines Setlist.fm, Spotify, FastAPI inference, speech-to-text, query understanding, transformer classifier experiments, and evaluation dashboards to make music intent observable and useful.
FastAPISTTTransformer classifierEvalsML InferenceInngestNext.js
Visit projectTime Crisis Wiki
A search and retrieval app for Time Crisis episode data, backed by a Go scraper, MongoDB Atlas, OpenAI embeddings, vector search, and a React interface for episode search and AI-assisted exploration.
GoMongoDB AtlasVector SearchEmbeddingsRAGData Pipeline
Visit projectPersonal Portfolio
This site: a Next.js portfolio with themeable design tokens, a terminal-inspired interaction layer, and content shaped around applied AI, developer tooling, and product craft.
Next.jsTerminal UIInteractive UIDesign Tokens
Visit projectCurrent Focus
What I am building and learning now.
Designing
AI architecture at DTN
Designing secure, identity-aware AI platform patterns at DTN: AWS Bedrock AgentCore Runtime, Strands agents, FastAPI MCP services, runtime boundaries, and self-service paths for product teams.
AWS Bedrock AgentCoreStrands AgentsFastAPI MCPIdentity/AuthAgent Runtime
Building
Music AI product loops
Using Wavves - Music Lab to build concert playlist generation, song recognition, speech-to-text, query understanding, eval baselines, and practical ML service boundaries.
STTDeepgramFastAPISpotifyEvals
Studying
Transformer fundamentals
Building intuition for tokenization, self-attention, multi-head attention, BERT-style classifiers, model artifacts, and evaluation loops.
TokenizationSelf-attentionBERT-style modelsModel artifactsEvals
Contact
Start the right conversation.
Want to talk applied AI platforms, developer tooling, product engineering, or a role where those overlap? LinkedIn is the best place to start. GitHub has more of the technical trail, and email is there when direct makes more sense.