Python Virtual Environments, Conda and Poetry, and Everyday Linux Commands

Explain what Python virtual environments are and why you need them, where tools such as Conda and Poetry fit in, and how you work on Linux day to day. These questions closed the fundamentals half of a technical screen for a junior quantitative developer role. Several were phrased as experience questions ("Have you used Conda?", "Do you use Windows or Linux more?"), so answer with what each tool is for and when you would choose it, backed by what you actually use. ### Clarifying Questions - Do the team's jobs run on Linux servers, and what do developers work on locally? - Does the team standardize on one environment tool, or does each project choose its own? - Do the projects depend on compiled non-Python libraries, such as numerical libraries or GPU toolkits? ### Part 1 — Python environments 1. What is a virtual environment? 2. Why do you need one? 3. Have you used Conda? What is it for? 4. Have you used Poetry or a similar tool? What does it add? ```hint Two projects, one machine Picture two projects on one machine that need different versions of the same library. Then picture handing one of them to a colleague, or deploying it to a server, and getting exactly the same versions there. ``` #### What This Part Should Cover - What a virtual environment contains and what activation actually does - The problems it solves: conflicting dependencies, reproducibility, isolation from the system interpreter - How Conda and Poetry differ from `venv` with `pip`, and when to choose each - Pinning and lock files for reproducible installs ### Part 2 — Working on Linux 1. Do you use Linux regularly? 2. Do you work more on Windows or on Linux? 3. Which Unix or Linux commands do you use often? ```hint A job that did not finish Think through what you would actually type on a remote server to find a failed job's log, search it for the error, and check whether the process is still running or the disk is full. ``` #### What This Part Should Cover - An honest description of your own working environment - Commands grouped by task: files, text search, processes, resources, remote work - Composing commands with pipes to answer a concrete question ### What a Strong Answer Covers - A precise explanation of how a virtual environment isolates packages, beyond "it keeps things separate" - Clear criteria for choosing `venv` with `pip`, Conda or Poetry, plus lock files for reproducibility - Fluency with the Linux commands needed to debug a job on a server, shown with a realistic sequence - Honesty about experience, combined with evidence of being able to work on Linux production machines - Awareness of the step beyond virtual environments, containers, for system-level reproducibility ### Follow-up Questions - A model runs on your laptop but fails on a Linux server with an import error. How do you find the cause, and how do you prevent it from recurring? - When would you choose Conda over `pip` with `venv`, and what goes wrong if you mix Conda and pip installs carelessly? - A nightly data job on a server did not produce its output. Which commands do you run, in what order? - How do you keep a long-running process alive after you disconnect from an SSH session?