← AT&T Interview Insights

AT&T·Software Engineer·Onsite - System Design / Architecture·Senior

Senior
May 2026

Summary

System design round focused on building an LLM-based log parsing pipeline. The problem was more open-ended than I expected and required thinking through schema variability across log types, which I hadn't really prepped for.

Questions Asked (1)

Q1

Design an LLM-based workflow that can parse and extract structured fields from log messages, where different log types (e.g. access logs vs. error logs) have completely different schemas.

System DesignTechnical Trade-offsAPI & Integrations
Author's notes

The tricky part isn't the LLM call itself, it's the schema variability.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying requirements and constraints, then propose a modular pipeline that classifies log types and routes them to schema-specific extraction prompts or models. Emphasize validation, fallback mechanisms, and scalability while discussing trade-offs between accuracy, cost, and latency.

Pro tip: Mention that you would use a two-stage approach: first classify the log type with a lightweight model, then use a specialized extraction prompt for that schema, and always include a validation layer to catch hallucinations and fall back to regex or human review.

1. Clarify Requirements and Constraints

Ask about log volume, latency requirements, schema variability, and accuracy expectations. Determine if real-time or batch processing is needed and what downstream systems consume the extracted fields.

2. Design a Modular Pipeline

Propose a pipeline with stages: ingestion, log type classification, schema-specific extraction, validation, and output. Use a router to direct logs to the appropriate extraction component based on classified type.

3. Choose LLM Strategies for Extraction

For each log type, design a prompt that includes the schema definition and few-shot examples. Consider using function calling or JSON mode to enforce structured output. Discuss fine-tuning vs. prompt engineering trade-offs.

4. Implement Validation and Fallbacks

Add a validation layer to check extracted fields against expected types, ranges, and patterns. If validation fails, fall back to rule-based parsing or flag for human review to maintain reliability.

5. Address Scalability and Cost

Discuss batching, caching, and using smaller models for classification. Consider asynchronous processing and monitoring for drift. Mention cost optimization by routing only complex logs to LLMs.

Key Points to Mention

  • Log type classification using a lightweight model or heuristics before extraction
  • Schema-specific prompts with few-shot examples and structured output enforcement (e.g., JSON mode, function calling)
  • Validation layer to detect and handle hallucinations or format errors
  • Fallback mechanisms to rule-based parsing or human-in-the-loop for low-confidence extractions
  • Scalability considerations: batching, caching, asynchronous processing, and cost management
  • Monitoring and continuous improvement: tracking accuracy, drift detection, and retraining/fine-tuning as needed

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.