Especialista de Engenharia de Software, IA

This is a fully remote position, open to applicants in Brazil.

📋 Description

• Lead the engineering strategy for AI solutions and platform (architecture, standards, technical roadmaps, and adoption guidelines).

• Define and maintain standards and best practices for the development, testing, and operation of AI applications (APIs, SDKs, templates, libraries, and accelerators).

• Monitor and drive the technical evolution of teams (architecture/code reviews, decisions, and trade-offs).

• Address complex problems in distributed environments.

• Serve as a technical reference in applied AI engineering: model serving, integration with LLMs, RAG, data/features pipelines, governance, and security.

• Track and improve engineering and AI metrics such as: SLO/SLI, latency, error rate, cost per request, response quality, hallucination, groundedness, test coverage, and incident rate.

• Collaborate on the design and development of platform tools and services focusing on security, performance, and observability (logs/metrics/traces, distributed tracing, end-to-end correlation).

• Ensure production operation and support: monitoring and alerts, participation in war rooms, post-incident (RCA), continuous improvements, and risk management.

• Guarantee scalability, resilience, and consistency of AI solutions: rate limit controls, fallback/degradation, caching, prompt/model versioning, data security, and guardrails.

• Continuously contribute ideas and proposals to enhance the roadmap (cost/latency reduction, quality improvement, evaluation automation, integration standardization).

⛳️ Requirements

• Solid experience as a Specialist/Staff/Tech Lead in software engineering and distributed systems architecture, with a strong cross-functional focus on AI.

• Mastery of microservices architectures and service integration (REST/gRPC, event-driven), with mature practices in CI/CD, testing, and versioning.

• Advanced experience in observability and troubleshooting (log analysis, distributed tracing like OpenTelemetry/Jaeger, APM), addressing latency issues, resource contention, and intermittent failures.

• Experience with messaging/streaming systems (e.g., Kafka) in production: designing topics/partitions, delivery guarantees, fault tolerance, idempotency, and throughput.

• Strong knowledge of design patterns and architectural principles, with the ability to apply them pragmatically.

• Experience with SQL and/or NoSQL databases, modeling, and optimization (tuning, indexes, queries), and the use of caching (e.g., Redis) for performance enhancement.

• Practical experience in building and operating solutions with LLMs/ML: serving design, integration via APIs/SDKs, latency/cost control, and fallback strategies.

• Experience with RAG and search (ingestion pipelines, chunking, embeddings, re-ranking, filters, and quality evaluation) and/or vector databases.

• Knowledge of MLOps/LLMOps practices: versioning and traceability, environments, experimentation, automated evaluation, drift/quality monitoring, and governance.

• Ability to define quality and security criteria for AI (prompt/flow testing, offline/online evaluations, hallucination mitigation, content policies, protection against prompt injection, and data leakage).

🏝️ Benefits

• 14th and 15th fixed salaries;

• Profit-sharing (based on seniority);

• Health and Dental plans with no co-payment;

• Well-being programs with Wellhub (formerly Gympass), Nutrition, Psychology, Occupational Health, Massage, running group, and local gym;

• Meal and Food Vouchers – flexible allocation between meal/food cards, no co-payment;

• Extended maternity and paternity leave;

• Childcare or nanny allowance for children up to 6 years and 11 months;

• Allowance for children with disabilities, with no age limit;

• Life insurance;

• Private pension contribution up to 8% of salary;

• Training platform – Sicredi Aprende, offering various courses;

• 40-hour workweek – utilizing a time bank system;

• Remote work allowance (except for positions that are 100% on-site).