MLOps.community

Software Engineering in the Age of Coding Agents: Testing, Evals, and Shipping Safely at Scale

70 snips
Feb 10, 2026
Guest
Ereli Eran
Ereli Eran, founding engineer at 7AI who builds agentic AI systems for security ops, joins to unpack real-world agent engineering. He covers how agentic systems mix deterministic code with stochastic LLM behavior. They talk testing, evals, safety gates, progressive prompts, model hybrids, observability and audit trails, and strategies for shipping agents reliably at scale.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

Hybrid Nature Of Agentic Systems

  • Agentic systems blend data-science uncertainty with traditional software engineering determinism.
  • You must design around stochastic LLM behavior to achieve predictable production outcomes.
ADVICE

Limit Agent Responsibility

  • Narrow agent responsibilities and inject context sparingly to avoid context pollution.
  • Use targeted instructions and guardrails rather than dumping all rules into every prompt.
ADVICE

Use A Constellation Of Models

  • Use multiple models or sub-agents for different tasks instead of one monolithic LLM.
  • Assign cheap fast models to trivial tasks and stronger models for reasoning to save cost and increase velocity.
Get the Snipd Podcast app to discover more snips from this episode
Get the app