TL;DR
Senior AI Engineer: Designing, optimizing, and automating LLM inference pipelines, ensuring scalability, performance, and cost efficiency with an accent on integrating testing and evaluation processes. Focus on profiling and optimizing model performance, automating LLM testing, and collaborating with AI QA and prompt engineers.
Location: Hybrid (Armenia)
Company
hirify.global develops AI-powered solutions to enhance audio communication, focusing on noise cancellation technology.
What you will do
- Design, optimize, and document end-to-end LLM inference pipelines for production, ensuring scalability and reliability.
- Profile and optimize model performance (speed, cost, compute) using cloud-based AI services (AWS, GCP, Azure).
- Monitor and log model performance, defining and implementing evaluation metrics for accuracy, latency, and cost efficiency.
- Automate LLM testing, including hallucination detection, bias monitoring, and robustness checks.
- Collaborate with AI QA and Prompt Engineers to integrate testing, evaluation, and prompt design into the pipeline.
Requirements
- Must be based in Armenia for hybrid work.
- Strong Python and ML framework expertise (PyTorch, Hugging Face).
- Deep understanding of LLMs (GPT, Claude, Mistral, etc.) and prompt engineering methodologies.
- Experience with vector databases and retrieval-augmented generation (RAG).
- Experience in profiling and optimizing LLM performance (latency, cost, memory usage).
- Strong knowledge of APIs and cloud AI infrastructure.
- Ability to document pipeline architecture and workflows for cross-team collaboration.
Nice to have
- Familiarity with LLM optimization and deployment strategies.
Culture & Benefits
- Commitment to an inclusive and respectful work environment.
- Valuing diversity and prohibiting discrimination based on various protected characteristics.
- Fostering a culture of respect and empathy among all employees and contractors.
Hiring process
- Application via online form.
Будьте осторожны: если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →