AI Platform Tech Lead

Job description

# **About DEUNA** DEUNA is a payments infrastructure platform that helps enterprise merchants across Latin America, the US, and Europe optimize and orchestrate their entire payment stack. We combine payment routing intelligence, AI-driven optimization, and a composable checkout experience to help companies increase revenue and reduce operational complexity at scale. We are backed by leading investors and processing billions of dollars in annual transaction volume. # **About the Role** DEUNA is a payments infrastructure company powering enterprise commerce across Latin America, the US, and Europe. We operate at the intersection of high-volume payment orchestration and applied AI — building intelligent systems that optimize authorization rates, reduce costs, and automate complex payment workflows for some of the largest merchants in the world. We are looking for a Staff/Principal-level AI Platform Tech Lead to own the full technical stack behind our AI payment intelligence and digital workforce products — from ML model training through production routing integration. This is a hands-on leadership role: you will set the architecture, write the code, and grow the team. # **What You Will Do** ## **ML & AI Systems** - Design, train, and own the full lifecycle of ML models for payment optimization — routing decisions, authorization rate improvement, cost reduction, and fraud signals — using PyTorch, TensorFlow, or XGBoost. - Build and operate LLM-powered workflows: LangGraph agent orchestration, RAG pipelines, and vector DB integrations (Pinecone, pgvector, or Weaviate). - Own the MLOps stack end-to-end: experiment tracking (MLflow / W&B), model registry, feature store, and automated retraining pipelines on AWS SageMaker. - Monitor model health continuously — drift, distribution shifts, retraining triggers — and define evaluation metrics tied directly to business outcomes. ## **Platform Engineering & Payments Integration** - Build and maintain inference services in Go and Python integrated into live payment routing — strict latency SLAs (<100 ms), zero silent errors. - Own AWS infrastructure: ECS/EKS, Terraform IaC, SQS/SNS event streaming, RDS/Aurora, and S3 for model artifacts. - Design and ship on-premise and hybrid deployment architectures for enterprise clients requiring local data residency, including secure data sync pipelines. - Apply PCI-DSS standards across all components touching payment data; implement tokenization in ML pipelines; design for PSP-specific behavior (Cybersource, Worldpay, Prosa, Cielo, Pagbank, and others). - Build and maintain RESTful and gRPC APIs that expose AI platform capabilities to merchants and partners. ## **Technical Leadership** - Own observability end-to-end: Prometheus/Grafana dashboards, OpenTelemetry tracing, model-specific monitors, and on-call runbooks. - Set the engineering bar for the team: architecture reviews, code standards, testing strategy (unit, integration, shadow mode), and CI/CD practices. - Mentor engineers, run design reviews, and translate product vision into executable technical roadmaps with clear timelines and trade-offs. # **Technical Skills** | | | | | | --- | --- | --- | --- | | **Backend / Platform** - Go (production services) - Python (ML + tooling) - gRPC & REST APIs - Event streaming (SQS/SNS) - Distributed systems | **Cloud & Infra — AWS** - ECS / EKS - Terraform / IaC - SageMaker or Vertex AI - RDS/Aurora, S3 - Hybrid / on-prem deploy | **AI / ML Stack** - PyTorch or TensorFlow - XGBoost / scikit-learn - MLflow / W&B - Feature stores - Model monitoring & drift | **LLMs & Agents** - LangGraph / LangChain - RAG + vector DBs - Prompt engineering - LLM evaluation - Structured outputs | | | | | | | --- | --- | --- | --- | | **Payments Domain** - PCI-DSS compliance - Tokenization patterns - PSP integrations - Auth rate optimization - Routing orchestration | **Frontend** - React / Next.js - TypeScript - Component systems - API integration | **Observability** - Prometheus / Grafana - OpenTelemetry - Structured logging - On-call runbooks | **Data** - SQL (analytical) - Airflow / dbt - Feature pipelines - Data quality & lineage | # **What We Are Looking For** - 8+ years in software engineering; 3+ at Staff, Principal, or Tech Lead level owning a production platform end-to-end. - Proven track record shipping ML/AI systems to production: training, serving, monitoring, and retraining — not just prototyping. - Hands-on LLM experience in production: agents, RAG pipelines, or AI workflow orchestration. - Payments or fintech background with practical knowledge of PSP behavior, PCI-DSS scope, authorization logic, and routing trade-offs. - Experience designing and deploying on-premise or hybrid enterprise infrastructure. - Bachelor's degree in Computer Science, Engineering, or equivalent demonstrated depth. # **What we offer** - A greenfield opportunity to define architecture, tooling, and engineering standards for an AI platform operating at scale across LatAm, US, and Europe. - Ownership of one of the most technically complex and business-critical systems at DEUNA — from model training through live payment routing. - Direct collaboration with product, operations, and modeling leadership — short feedback loops, high autonomy, real impact. - Competitive compensation, hybrid work and a team that takes engineering craft seriously.