instalilyAI·AI Engineer·Onsite - Coding / Algorithms·Intermediate

Intermediate

Jun 2026

Summary

Onsite live build round at instalilyAI where you have about four hours to ship a working fullstack AI app from scratch, including web scraping and AI-driven processing on the scraped data. Less about algorithms, more about whether you can actually deliver something end-to-end under real time pressure.

Questions Asked (1)

Build a complete fullstack AI application from scratch in ~4 hours. It should include web scraping, AI-driven processing of the scraped data, and a working frontend plus backend.

System DesignAPI & IntegrationsTechnical Trade-offs

Author's notes

Four hours sounds like a lot until you're twenty minutes in and still arguing with yourself about folder structure.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the scope and constraints with the interviewer, then propose a minimal viable architecture that prioritizes a working end-to-end flow over perfection. Focus on selecting simple, reliable tools and libraries for each component (scraping, AI processing, frontend, backend) and describe how you would integrate them within the time limit.

Pro tip: Emphasize that you would build a thin vertical slice first—scrape one page, process it with a simple AI model, and display results—then iterate. This demonstrates pragmatism and risk management, which interviewers value over ambitious but incomplete designs.

1. Clarify requirements and constraints

Ask about the target website, expected data volume, AI task specifics (e.g., summarization, classification), and deployment environment. Confirm that a minimal working prototype is acceptable and identify any must-have features.

2. Design a minimal architecture

Propose a simple stack: Python with BeautifulSoup/Scrapy for scraping, FastAPI for backend, a pre-trained model or API (e.g., OpenAI) for AI processing, and React or plain HTML/JS for frontend. Ensure components can be built and integrated quickly.

3. Plan the implementation timeline

Allocate time: 1 hour for scraping and data storage, 1 hour for AI integration, 1 hour for backend API, 1 hour for frontend and integration. Include buffer for debugging and deployment.

4. Address potential challenges and trade-offs

Discuss handling scraping blocks (e.g., rate limiting, user-agent rotation), AI model latency/cost, and data consistency. Explain how you would simplify or mock components if time runs short.

5. Summarize and validate

Recap the end-to-end flow, highlight that the solution meets the core requirements, and mention how you would test and demo it. Invite feedback or questions.

Key Points to Mention

Choice of scraping library (e.g., BeautifulSoup, Scrapy) and handling anti-scraping measures
AI processing approach: using pre-trained models vs. APIs, and managing latency/cost
Backend framework (e.g., FastAPI, Flask) and API design for frontend consumption
Frontend simplicity: using lightweight frameworks or vanilla JS to save time
Data flow and storage: temporary storage (e.g., SQLite, in-memory) for quick setup
Time management and prioritization: building a vertical slice first, then iterating

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