Forward Deployed Engineer, Brazil

The Role

We're looking for a Forward Deployed Engineer to join our LATAM team, with a focus on São Paulo. In this remote, customer-facing role, you'll embed directly with enterprise customers across the region to architect and ship production systems on Telnyx's global network — voice, messaging, AI, and wireless — including self-hosted, open-weight LLM deployments that run inside customer environments. This isn't about demoing products; it's about building real solutions that work at scale.

This role would suit a technical professional who thrives at the intersection of engineering and customer success. You're equally comfortable debugging SIP traces as you are standing up a LiteLLM gateway with routing and failover across a customer's model providers, or leading a whiteboard session with a customer's engineering team. The ability to own outcomes from POC to production go-live is essential.

Customers across Latin America increasingly need Spanish- and Portuguese-first conversational AI, regional data residency, and flexible deployment options. A significant part of this job is making open-weight models perform reliably on infrastructure the customer controls. Professional fluency in Spanish or Portuguese, together with strong working proficiency in English, is required.

This is a LATAM-based remote role for candidates located in São Paulo, with approximately 10–30% travel for deployments, workshops, and escalations across Brazil, and the wider region.

Responsibilities

• Embed with enterprise customers to understand their communications workflows, AI use cases, and integration challenges firsthand
• Build and deploy custom implementations: AI Voice Assistants, Telnyx APIs (Voice, Messaging, Fax, Wireless), and WebRTC
• Deploy and operate open-weight LLMs such as Llama, Qwen, Mistral, DeepSeek, and gpt-oss in customer-controlled cloud, on-premises, and air-gapped environments
• Deploy and operate LiteLLM as the model gateway in customer environments, providing a unified OpenAI-compatible interface across self-hosted models and hosted providers, with routing, load balancing, retries and fallbacks, rate limits, and per-team virtual keys
• Instrument and govern LLM usage through cost tracking and budgets, caching, logging and observability using OpenTelemetry, Langfuse, or similar tools, and guardrails
• Design model-routing strategies for real-time voice workloads, balancing latency, cost, reliability, and Spanish- or Portuguese-language quality
• Make the build-vs-buy case between self-hosted open-weight models and hosted frontier APIs, while keeping customer application code portable across both
• Adapt models to customer domains through prompt and RAG pipelines, LoRA/QLoRA fine-tuning, and evaluation harnesses for Spanish, Portuguese, and multilingual use cases
• Lead POCs, pilots, and production launches from whiteboard to go-live
• Own customer outcomes and remain engaged until the solution is live and stable
• Collaborate directly with Product and Engineering to shape the roadmap based on field insights from the Latin American market
• Create clear technical documentation, runbooks, and maintainable solutions for handoff in English and, where needed, Spanish or Portuguese
• Troubleshoot and resolve complex integration issues alongside customer teams
• Help customers address applicable privacy, security, telecommunications, and data-residency requirements; including Brazil's LGPD

What We Are Looking For

• CS degree or equivalent experience
• 3+ years building software or doing technical consulting
• Proficiency in multiple languages such as Python, Node.js, and Go — you're more dangerous in some than others
• Hands-on experience running LiteLLM or a comparable LLM gateway such as Portkey, Kong AI Gateway, or an in-house proxy in production, including model definitions, proxy configuration, routing and fallback rules, and virtual key management
• Practical understanding of production model-gateway challenges: provider rate limits and quotas, timeouts and retries, streaming, token accounting and cost attribution, and concurrency-related failure modes
• Comfort deploying containerized services on Kubernetes, with secrets management, configuration, observability, and upgrades as part of the deployment story
• Strong API fluency, event-driven thinking, and cloud-native instincts
• Exposure to SIP, WebRTC, VoIP, or real-time voice and messaging systems
• Ability to translate “it's not working” into root cause
• Comfort working directly with customers and participating in high-stakes technical conversations
• Professional fluency in Spanish or Portuguese
• Professional working proficiency in English for internal collaboration, documentation, and work with Product and Engineering
• Based in or near Bogotá, Mexico City, or São Paulo, with authorization to work in the applicable country
• Willingness to travel approximately 10–30% across Latin America

Bonus Points For

• Professional proficiency in both Spanish and Portuguese
• Experience with AI voice assistants, STT/TTS, or LLM-based conversational systems, particularly for Latin American Spanish or Brazilian Portuguese
• Fine-tuning and post-training experience: LoRA/QLoRA, distillation, preference tuning, or building evaluation sets for a specific domain
• Experience with an inference-serving engine behind the gateway such as vLLM, SGLang, TGI, or Ollama, including sizing GPU capacity for self-hosted models
• SQL proficiency with PostgreSQL, MySQL, or Oracle
• ETL and data-wrangling experience
• DevOps fundamentals including Docker, Kubernetes, and CI/CD
• Background in telecom, CPaaS, contact centers, or high-growth SaaS
• Experience designing private-cloud, sovereign-cloud, on-premises, or data-residency-sensitive architectures
• Security mindset, including IAM, encryption, secrets management, and audit logging

#LI-RH1