Predict future time-series values

Quick Overview

This interview question evaluates ML product requirements, data/labeling, modeling, serving architecture, evaluation, monitoring, and trade-offs in a realistic interview setting. A strong answer for Predict future time-series values states assumptions, handles edge cases, explains trade-offs, and shows how to validate the result clearly.

Predict future time-series values

Company: Jane Street

Role: Machine Learning Engineer

Category: ML System Design

Difficulty: hard

Interview Round: Technical Screen

##### Question Given a time-series dataset, build an end-to-end machine-learning solution that forecasts future values: choose an appropriate model architecture, implement the training loop (e.g., in PyTorch), and explain your design decisions and evaluation strategy.

Quick Answer: This interview question evaluates ML product requirements, data/labeling, modeling, serving architecture, evaluation, monitoring, and trade-offs in a realistic interview setting. A strong answer for Predict future time-series values states assumptions, handles edge cases, explains trade-offs, and shows how to validate the result clearly.

|Home/ML System Design/Jane Street
Jane Street logo
Jane Street
Aug 4, 2025, 10:55 AM
hardMachine Learning EngineerTechnical ScreenML System Design
24
0

Predict future time-series values

End-to-End Time-Series Forecasting (PyTorch)

Context

You are given one or more regularly sampled numeric time series and optional exogenous covariates (calendar features, prices of related assets, etc.). The goal is to forecast the next H time steps.

Assume:

  • Input time steps are ordered and may contain missing values/outliers.
  • Forecast horizon H and input context length L are to be chosen by you.
  • You should avoid data leakage and use time-aware validation.

Task

Design and implement an end-to-end forecasting solution:

  1. Choose and justify a model architecture suitable for multi-step forecasting.
  2. Implement the training loop in PyTorch, including dataset/windowing and evaluation.
  3. Explain data preparation, hyperparameters, and design decisions.
  4. Describe an evaluation strategy (splits, baselines, metrics, and backtesting).

Clarifying Questions to Ask Guidance

  • Clarify users, core use cases, read/write patterns, scale, latency, availability, and data retention.
  • State explicit assumptions before making sizing or architecture decisions.
  • Prioritize the functional path first, then address reliability, security, observability, and rollout.

What a Strong Answer Covers Guidance

  • A scoped requirements summary with concrete non-goals and success metrics.
  • ML-specific data, model, evaluation, serving, and monitoring choices.
  • Reasoned trade-offs among simple and scalable designs, including bottlenecks and failure modes.
  • A validation, monitoring, migration, and launch plan appropriate for the risk level.

Follow-up Questions Guidance

  • What breaks first at 10x traffic or data volume?
  • How would you degrade gracefully during dependency failures?
  • What metrics and alerts would prove the design is healthy after launch?

Submit Your Answer to Earn 20XP

Sign in to leave a comment

Loading comments...

Browse More Questions

More ML System Design•More Jane Street•More Machine Learning Engineer•Jane Street Machine Learning Engineer•Jane Street ML System Design•Machine Learning Engineer ML System Design