How to apply for Software Engineer (Backend-Focused)

About AZX, PBC.

AZX, PBC. is a profitable, bootstrapped public benefit corporation founded in 2024, focused on AI transformation in critical industries like clean energy, real estate, and logistics. They work with category leaders and are building a long-term company for those passionate about AI and positive impact.

About the role

This role is the technical backbone of AZX's inference and evaluation systems, involving GPU scheduling, autoscaling, and serving infrastructure for vLLM/SGLang, as well as building evaluation systems to measure model and prompt improvements. It's high-leverage and offers architectural ownership over hard ML infrastructure problems.

A typical day

A typical day might involve designing a new autoscaling policy for GPU clusters, implementing a service to evaluate model outputs, or collaborating with data scientists to optimize prompt performance. You'd also be monitoring system health, writing code, and participating in technical discussions with the team.

Who AZX, PBC. is looking for

• Experienced with serving frameworks like vLLM or SGLang and GPU cluster management (e.g., Kubernetes, Ray).
• Strong background in backend engineering, with proficiency in Python and experience building scalable, low-latency APIs.
• Deep understanding of ML model evaluation, including offline and online metrics, A/B testing, and prompt engineering.
• Comfortable with ambiguity and able to set technical direction in a fast-growing startup environment.

Tips for this application

• Tailor your resume to highlight specific projects involving vLLM/SGLang, GPU optimization, or large-scale inference systems, not just generic backend work.
• In your cover letter, mention how your experience aligns with AZX's mission of positive impact in clean energy or other critical industries.
• Show tangible results: quantify improvements in latency, throughput, or cost reduction from your previous infrastructure work.
• If possible, contribute to or reference open-source projects related to inference serving or evaluation tools to demonstrate expertise.
• Research AZX's clients (CBRE, LevelTen Energy, Flexe) and mention how your work could help solve their industry-specific challenges.

What to cover in your cover letter

["Emphasize your passion for AI's potential to drive positive impact in critical industries like energy and climate.", 'Highlight your experience with large-scale ML systems, especially inference optimization and evaluation frameworks.', "Mention your ability to take ownership of complex infrastructure and set technical direction, aligning with the role's high-leverage nature.", "Show that you're aligned with AZX's values as a public benefit corporation and their focus on long-term success."]

Research before applying

• Read AZX's website and any public materials to understand their AI transformation approach and industries they serve.
• Look up their clients (CBRE, LevelTen Energy, Flexe) to understand the types of challenges they solve.
• Research public benefit corporation status and what it means for AZX's mission and operations.
• Familiarize yourself with the latest in inference serving (vLLM, SGLang) and evaluation frameworks (e.g., LangSmith, MLflow) to speak knowledgeably.

Likely interview topics

Based on the job description, expect questions about:

• How would you design a scalable inference platform for multiple models with autoscaling across cloud and customer-managed clusters?
• Describe your experience with vLLM or SGLang: what were the trade-offs and how did you optimize performance?
• How do you approach building evaluation systems for LLM outputs? What metrics and methods do you use?
• How would you handle GPU resource contention and scheduling in a multi-tenant environment?
• Walk me through a time you had to make architectural decisions for a critical system under tight deadlines.

Common mistakes to avoid

• Don't apply with a generic resume that doesn't highlight ML infrastructure or backend experience specifically.
• Avoid focusing only on model training; this role is about serving and evaluation, not training.
• Don't underestimate the importance of the company's mission; they likely value candidates who care about positive impact, not just technical challenges.