177: Vector Databases

Patrick and Jason discuss what it means to become a manager and how the role differs from individual engineering work. They cover hiring, coaching, performance management, team goals, and when moving into management is or is not the right choice.

Patrick and Jason break down workflow orchestrators and why they matter for batch jobs, long-running tasks, and resumable distributed systems. They compare tools such as Airflow, Dagster, Temporal, Ray, and Kubeflow while explaining the infrastructure patterns behind them.

Patrick and Jason explain asynchronous programming and how it differs from traditional multithreading and multiprocessing. They cover coroutines, blocking versus non-blocking operations, promises, callbacks, async/await, and the tradeoffs behind each approach.

Patrick and Jason are joined by Mark Cunningham to discuss how software engineers can find strong job opportunities and perform well throughout the interview process. They cover sourcing strategies, reverse interviews, negotiation, hiring-manager expectations, and common mistakes candidates should avoid.

Patrick and Jason discuss how AI-assisted coding tools can speed up development, answer questions about a codebase, and reduce boilerplate work. They compare common workflows and tools such as Copilot, Cursor, and command-line assistants while talking through where these systems help most.

Patrick and Jason cover memory management from both the operating-system and language-runtime perspectives. They discuss heap management, virtual memory, garbage collection, ownership models, and practical techniques for diagnosing and reducing excessive memory use.

Patrick and Jason introduce reinforcement learning and place it alongside supervised and unsupervised learning. They cover Q-learning, SARSA, policy gradients, actor-critic methods, PPO, imitation learning, and why training and evaluating RL systems is so challenging.

Patrick and Jason discuss project planning and management for software teams. They cover why planning matters, how frameworks like SMART goals, Gantt charts, Scrum, Agile, and Kanban fit together, and how to deal with uncertainty, dependencies, and scheduling risk.

Patrick and Jason revisit working from home and the realities of remote engineering work. They cover communication, scheduling, home-office setup, motivation, distractions, and why remote work is not equally effective for every team or every person.

Patrick and Jason explain vector databases by starting with embeddings, similarity metrics, and approximate nearest-neighbor search. They discuss how these systems store and query high-dimensional vectors and where tools like pgvector, Weaviate, Pinecone, and Milvus fit.