Cesar Augusto Ruiz Puentes Senior software engineer, AI and agentic systems

I build systems and products, end to end. Payment cores that move real money, and agentic AI of my own.

Agentic AI
MCP servers, persistent memory and context, provider integration, voice.
Distributed backends
Payment cores, payouts and FX on AWS, and the migrations that got them there.
End to end
Architecture decision, service, interface, pipeline and the deploy.
cr8297408@gmail.com +57 304 643 2942 GitHub LinkedIn

At a glance

Experience
4+ yrs
Own products
05
Client platforms
08
Base
Colombia
Work
Remote
Status
Open
Available 2026
4+
Years shipping and operating production software
~40%
Infrastructure cost cut migrating Xell off its monolith
07
Domains in the 88Pay core I led from zero
05 + 08
Products of my own, plus client platforms I built on

Selected work

The products I own
AI product, founder
2026
Founder and architect

NexusMind

A multi-tenant persistent memory and context platform, so a team's AI tools and agents share one auditable memory instead of seven private ones.

Decision Each request resolves API key to user to organisation before storage is touched, so tenancy lives in the query layer rather than in handler code where it can be forgotten.
Trade-off Cross-org features, like sharing a memory between teams, will need an explicit escape hatch later.
Open the case study4 decisions, 8 node diagram, API shapes
MCP
Architecture
600 × 380
Terminal
$ nexusmind --status

ok  services   running
ok  tests      passing
600 × 380
TypeScriptMCPRustAxumReactDocker
AI developer tooling
2026
Author

J.A.R.V.I.S.

A hands-free developer interaction system: a PTY wrapper, a streaming lexer, real-time voice activity detection and interchangeable STT, TTS and LLM adapters.

Decision Everything runs through interchangeable adapters and real-time streams, so no single model or speech vendor is baked into the system.
Trade-off An adapter layer for every provider is more surface to maintain than calling one API directly, and it is the only way to swap engines by config.
Open the case study3 decisions
Voice
1200 × 860J.A.R.V.I.S. primary view
Voice
Architecture
600 × 380
Terminal
$ jarvis --status

ok  services   running
ok  tests      passing
600 × 380
PythonPTYVADSTT / TTSLLM adapters
Company, co-founder
2024 to 2026
Co-founder and architect

SmartCoder Labs

An AI-assisted software engineering ecosystem: code generation and refactoring, automated testing, and delivery and quality controls wired into the same pipeline.

Decision CI/CD, SonarQube, Sentry and OpenTelemetry turn model output into measurable software-quality signals instead of unreviewed volume.
Trade-off Gates slow a single change down. Without them, AI-generated volume is exactly what nobody can review.
Open the case study2 decisions
AI SDLC
1200 × 860Quality gate on an agent-authored pull request
AI SDLC
Architecture
600 × 380
Terminal
$ smartcoder-labs --status

ok  services   running
ok  tests      passing
600 × 380
CI/CDSonarQubeSentryOpenTelemetryAgents
Fintech product, mobile
2025 to 2026
Author

Capitalia

A personal finance app with real subscription billing: budgets, accounts and goals on Supabase, and a payment layer that moved from Stripe to Wompi without the schema having to care which one it is.

Decision Moving to Wompi could have been a rename. Instead the customer table became payment_customers with an explicit payment_provider column, so the data model stopped encoding a vendor name.
Trade-off A rename migration against live tables is riskier than adding new ones, and two provider implementations now have to keep working rather than one.
Open the case study4 decisions, 6 node diagram
Mobile
1200 × 860App dashboard: budgets, accounts, goal progress
Mobile
Architecture
600 × 380
Terminal
$ capitalia --status

ok  services   running
ok  tests      passing
600 × 380
ExpoReact NativeTypeScriptSupabasePostgres RLSEdge FunctionsWompiStripe
Open source, npm package
2025 to 2026
Author

Apple Design System

A React design system following Apple human interface guidelines, shipped two ways at once: an npm package you depend on, and a registry CLI that copies components into your repo so you own them.

Decision Teams that want a dependency get one; teams that need to own and edit the component copy it in. The same source produces both, so neither is the neglected path.
Trade-off Two distribution mechanisms to keep in step, and a registry build that has to run on every release.
Open the case study3 decisions
DS
1200 × 860Apple Design System primary view
DS
Architecture
600 × 380
Terminal
$ design-system --status

ok  services   running
ok  tests      passing
600 × 380
React 19TypeScriptTailwind CSSRadix UIFramer MotionnpmCLI
Client platforms 8 projects, delivered for clients. Problem and solution only, no client architecture.
Portrait
1000 × 1150Portrait. Desk or whiteboard, half-lit profile.

How I work

“I take ownership of the whole thing: the architecture decision, the service, the interface and the deploy. If the trade-off cannot be stated in one sentence, the design is not finished.”

I work across backend, frontend, cloud and delivery, from the architecture decision to production, in TypeScript, Rust and Python. Every project below ships with its architecture, the decision behind it and the trade-off I accepted.

I own the whole slice

Interface, service, infrastructure and pipeline. Small teams do not have hand-off points, so I work as if there are none.

I design for failure

In payments a bug is an accounting problem, so validation, rate limiting, retries and observability are part of the first version.

I hand over cleanly

I write a system up the way I would want to receive one I am about to maintain: decisions, trade-offs, runbook.

Capabilities

What I get hired for

Agentic AI and tooling

I build the layer underneath AI products: MCP servers and tool protocols, persistent memory and context, provider integration, speech to text, text to speech and voice activity detection.

MCPLLM appsAI agentsSTT / TTSVADClaude Code

Full-stack product

I ship features end to end: React and Next.js interfaces on top of my own APIs, third-party integrations, webhooks and the data model behind them.

TypeScriptReactNext.jsTailwindNestJSPrismaGraphQL

Distributed backend and fintech

Payment cores, payouts, commissions and foreign exchange on AWS, plus the monolith-to-microservices migrations that got them there. In money systems a bug is an accounting problem.

Node.jsNestJSGoRustPythonEvent-drivenServerless

Cloud, delivery and reliability

I provision as code, test at every level and instrument for real questions, so the on-call answer is a query rather than a guess.

AWS LambdaEKSTerraformDockerGitHub ActionsOpenTelemetryJest

From decision to production

  1. I write the decision record before the first commit.

  2. I draw the architecture as data, so it never goes stale.

  3. Contracts and types get reviewed before implementation starts.

  4. Then I instrument it, so production answers questions.

Experience

2021 to now
Jan 2023 to present

Byte4Bit

Senior backend engineer
Remote

I own backend services, APIs and architecture decisions for production financial platforms: microservices, async workflows, cloud infrastructure, automated testing, CI/CD, observability and incident-oriented reliability, working alongside frontend, DevOps, QA and product while driving technical solutions from design to deployment.

88Pay

Payments core, AWS

Led the core from zero through a full microservices migration: payments, accounts, identity, files, analytics, notifications and public APIs. Node.js and TypeScript services processing real-money transactions with strict validation, rate limiting, secure coding and observability.

TWL / TWLPro

Backoffice, EKS

Migrated the platform to microservices on AWS with AppSync GraphQL, NestJS Lambdas, PostgreSQL, Cognito and Terraform, and helped build the orchestration layer that coordinated Puppeteer extractors on EKS and fed reporting, payroll, commission and payment pipelines.

Audit1

ACH payments, NACHA

Owned the architecture, implementation and deployment of the ACH payment processor that collects workers compensation premiums: NACHA file generation and reconciliation, returns and transaction-status handling, built hexagonally with the payment domain isolated from HTTP, storage and the message bus, and shipped with its own Terraform.

Xell

Migration, ~40% cheaper

Led the migration of a Node.js and TypeScript monolith to asynchronous AWS Lambda microservices, cutting infrastructure cost by roughly 40 percent while improving service boundaries; designed the DynamoDB and MySQL persistence models and the Jest unit, integration and end-to-end suites.

Jul 2022 to Jan 2023

Kabalaw

Backend engineer
Miami, FL, remote

Modernised Node.js backend services and redesigned APIs for scalability, modularity and maintainability without interrupting operations.

Module generators

Standardisation

Built reusable module generators and templates so the team could standardise architecture, accelerate delivery and keep the codebase evolving consistently.

2024 to present

SmartCoder Labs

Co-founder and software architect
Colombia, remote

I design an AI-assisted software engineering ecosystem: code generation and refactoring, automated testing and delivery and quality controls, combining agentic workflows with production engineering practice.

Quality pipeline

CI/CD, quality signals

Integrated CI/CD, SonarQube, Sentry and OpenTelemetry so agentic development produces measurable software-quality signals rather than unreviewed output.

Feb 2021

Bioscenter

Freelance backend developer
Colombia

Designed a point of sale with two independent APIs, centralised licensing and local operations for inventory, users and transactions, documented with Swagger.

Two-API split

Licensing, offline-first

Separating licensing from local operations let each store keep selling when connectivity dropped, without giving up central control of licences.

Education

Systems engineering

Universidad Nacional Abierta y a Distancia (UNAD)
2023 to present

Backend web development

Acamica / Protalento
Feb 2022 to Sep 2022

Need someone
who owns it
end to end?

I am open to senior product-engineering and AI roles, and to fixed-scope work: architecture reviews, monolith-to-microservices migrations, or an MCP server for your product. Reply within 24 hours.

cr8297408@gmail.com
GitHub LinkedIn