AI Technical Director

We are seeking a Director or Technical Leader with deep, hands-on experience in Generative AI to lead the creation and growth of our AI Center of Excellence in LATAM. This is a builder role, you will be the first of your kind in the region, shaping the structure, training the next generation of engineers, and enabling scalable GenAI delivery in collaboration with our North American (NAMER) AI team.

You will define the region's technical direction, establish best practices, train and onboard talent, and staff and lead cutting-edge GenAI projects across industries.

Core Responsibilities:

Technical Leadership

• Architect scalable, end-to-end GenAI solutions with real-world deployment experience.
• Drive technical vision and roadmap for GenAI services in LATAM
• Ensure a culture of continuous learning, engineering excellence, and responsible AI

Delivery & Execution

• Lead GenAI projects hands-on: from client pitch to architecture, implementation, and deployment
• Oversee projects that span frontend, backend, infrastructure, data pipelines, and AI components
• Advocate for automation, testing, monitoring, and CI/CD in AI pipelines

Team Development

• Develop talent pathways and create a pipeline of GenAI leaders in the region
• Conduct regular workshops, code reviews, and innovation labs
• Empower junior engineers through mentorship and clear development plans

Labs & Research

• Establish and lead internal AI labs focused on experimental ideas, rapid prototyping, and applied research initiatives.
• Continuously scan academic and industry literature to identify high-impact research opportunities relevant to company goals.
• Analyze and distill complex scientific papers, communicating key insights and implications to cross-functional teams.
• Define research directions, methodologies, and success criteria in collaboration with scientists and engineers.
• Lead the end-to-end lifecycle of research initiatives—from ideation and experimentation to validation and integration into production systems.

Qualifications:

• 10+ years in software engineering, including recent roles in Director-level or higher leadership
• 2+ years building or shipping production-grade GenAI solutions (text/image/video or multimodal)
• Deep technical knowledge of multiple traditional stacks:
• Backend (Node.js, Python, Go, etc.)
• Frontend (React, WebGL, etc.)
• Cloud & Infrastructure (AWS preferred)
• DevOps & CI/CD (Docker, GitHub Actions, Terraform, etc.)

• Proficient in training and fine-tuning frameworks such as Pytorch and HF Transformers
• Strong understanding of inference and deployment tooling (ONNX, Triton Inference Server, etc)
• Experienced with cloud-based AI infrastructure: AWS SageMaker, GCP Vertex AI, Azure ML, Kubernetes.
• Hands-On Experience in Production-Scale ML Systems

• Large Language Models (LLMs):
• Fine-tuning models like GPT, LLaMA, Mistral on domain-specific data.
• Building pipelines for prompt engineering, RAG (retrieval-augmented generation), and instruction tuning.
• Deploying models via APIs and local inference runtimes (vLLM, LoRA)
• Implementing and using evaluation frameworks to assess output quality (e.g., relevance, coherence, safety, bias) with clear metrics.
• Retriever quality assessment and optimization.
• Designing and operating scalable vector stores (e.g., Pinecone, OpenSearch), including chunking strategies and techniques for detecting underperforming retrieval setups.
• ETL pipelines for GenAI leveraging structured, semi-structured, and unstructured data sources.
• Monitoring, tracing, and debugging complex GenAI workflows in production environments.
• Generator quality improvement techniques such as reranking, voting, and consensus models.
• Responsible AI practices including PII handling, moderation, fairness, and model guardrails.
• Strong understanding of trade-offs between cost, latency, and quality in GenAI system design.
• Experience with data privacy, security, and observability in AI pipelines.
• Computer Vision (CV):
• Training image classification, object detection, and segmentation models
• Working with pretrained architectures (ResNet, EfficientNet, ViT)
• Deploying vision models in multimodal systems (e.g., captioning, OCR, vision-language tasks)
• Diffusion Models:
• Training and customizing generative diffusion models for production use
• Familiarity with latent diffusion architectures (e.g., Stable Diffusion) and noise schedulers (DDPM, DDIM, etc.)
• Deploying fine-tuned models and scaling inference across multiple regions and environments.
• Monitoring model behavior in production, including output consistency, performance degradation, and error tracking.
• Translating complex workflows (e.g., ComfyUI, Automatic1111) into highly optimized, maintainable production code.
• Integrating diffusion workflows into creative pipelines and user-facing applications

• Expert in quantitative and qualitative evaluation methodologies
• Degree in Computer Science, Engineering, or equivalent hands-on experience
• Excellent spoken and written English
• Strong communication and mentoring capabilities

#LI-LT1 #LI-Hybrid