Senior Web Scraping Engineer

Senior Web Scraping Engineer (Python) — India (Remote)

Employment type: Full-time (open to contract-to-hire)

Work location: Remote in India

Time overlap: Prefer 2–3 hours/day with Pacific Time (PST)

About the role

We're building a high-throughput product data ingestion pipeline across hundreds of domains. You'll own the crawling/extraction layer end‑to‑end: crawling with a Playwright fallback, per‑domain learned selectors, and reliable PDF handling (datasheets/specs). You'll also drive the automation around scheduling, retries, and monitoring so runs are hands‑off, and you'll integrate vendor/public APIs (REST/GraphQL) wherever available to complement crawling.

This role spans crawling (discovering & fetching pages via sitemaps/robots) and scraping (extracting structured specs, images, and PDFs into our schema).What you'll do

• Design an crawler (Scrapy or aio with Playwright fallback only for JS‑heavy pages.
• Implement sitemap diffing and conditional GETs (ETag/Last-Modified) for incremental runs.
• Build a lightweight "needs JS?" classifier (HTML length, JSON‑LD presence, data‑product markers) to auto‑route vs Playwright.
• Enforce per-domain throttles/backoff (2–4 concurrent/domain; auto‑lower on 429/503).
• Add URL normalization/canonicalization and de‑dup (respect ; hash PDFs).
• Handle PDF discovery & download (HEAD first to dedupe; size/concurrency caps; SHA‑256 keys).
• Apply Playwright browser automation resource budgets (block images/fonts/analytics; kill outliers by size/CPU/time).
• Integrate third‑party APIs (REST/GraphQL) as first‑class sources: handle auth (API keys/OAuth2), pagination, and rate limits; unify API + crawl outputs .
• Own automation & orchestration for scheduled runs (Airflow/Temporal/Celery or cron), idempotent retries, and alerting.
• Create per‑domain selectors (YAML) with verification on hold‑outs; re‑learn only when health drops.
• Ship observability : per‑site field coverage, error rates, retries, avg page time, and PDF success.
• Maintain allow/deny paths ; adhere to and Terms of Service.
• Containerize workers; provide runbooks/CI; collaborate with data team on schemas/normalization.

Must‑have qualifications

• 4+ years Python , including 2+ years building production web crawlers at scale.
• Strong with Scrapy or aio and Playwright (or Puppeteer) in production.
• Practical proxy management , polite anti‑bot tactics, and per‑domain rate limiting .
• Hands‑on with ETag/Last-Modified , retries, backoff, and caching.
• Confident with CSS/XPath , ‑LD , and HTML parsing.
• APIs: consuming REST/GraphQL (auth, pagination, backoff) and building small internal services ( FastAPI or similar).
• Automation/Orchestration: Airflow/Temporal/Celery (or equivalent schedulers/queues) for scheduled runs and monitoring.
• PDF handling (requests/HEAD, hashing, size limits) and file integrity checks.
• Queues ( Redis/Kafka ), Docker , Linux basics; comfort with logs/metrics.
• Clear, pragmatic communication and strong ownership.

Nice to have

• Go or experience for high‑performance crawlers.
• Cloud: AWS/GCP , S3 , ECS/Kubernetes; IaC basics.
• Workflow engines: Airflow/Temporal/Argo/Celery .
• Document extraction: Textract/Tika/Camelot/Tabula .
• Search/analytics: Elasticsearch/OpenSearch ; warehousing ( Snowflake/Postgres ).
• LLM‑assisted selector generation with deterministic verification (optional).

How we work

• Ship in small, measurable increments.
• Track coverage and freshness as north‑star metrics.
• Prefer simple designs that are easy to operate at scale.

Compensation

Competitive; please include your expected CTC (INR LPA) and any variable/benefits expectations.

Application

Please apply with your resume and links to relevant repos or code samples. Include concise notes on:

1. a crawler you ran at 100+ sites/day (or similar scale),
2. how you handled rate limits/retries , and
3. your approach to PDF discovery/dedup .
   

---