Engineering deep-dives on the work I actually do — production scraping, browser automation, AI extraction pipelines, and the back-office systems that hold them together. Each article ties back to a topic hub and, where it applies, a real case study.
How to deduplicate company records scraped from dozens of sources into one canonical row per real-world company — the match ladder, fill-blanks-only merging, provenance rows that never get destroyed, and the counting bugs that bite everyone who builds one of these.
Read the article →A production architecture for B2B contact enrichment — free website extraction first, credit-gated API reveals second, premium sources last, with score gating, no-double-spend guards, and upward-only verification status so every credit lands on a company worth calling.
Read the article →How to detect purchase intent from what companies do in public — job postings across 12 ATS platforms, LLM-classified news monitoring, customs import records, and domain-health checks — and why signals should rank your call queue but never touch your lead score.
Read the article →How I built a production OCR extraction pipeline that reads a paywalled digital-flipbook trade directory page by page and turns it into clean, auditable company records — segmentation strategy, strict-vs-loose parsing, an LLM fallback with a deletion guard, and the verification discipline that caught every real defect.
Read the article →A walkthrough of how I structure production Python scrapers — pipeline stages, idempotent storage, block-aware fetching, and the architectural decisions that separate a one-shot script from a system you can put on a schedule.
Read the article →A practical comparison of Playwright and Selenium for production browser automation work — selector engines, async control, network interception, anti-detect integration, and why I default to Playwright for ~95% of new projects.
Read the article →How to architect a parallel browser automation worker pool with Python, Playwright, and MySQL — queue design, worker partitioning, recovery from crashes, error capture, and the operational discipline that keeps long-running runs stable.
Read the article →How to drive Kameleo browser profiles from Python and Playwright over CDP — profile lifecycle, proxy-per-profile pairing, parallel worker pools, and the operational patterns that make Kameleo viable for long-running scraping work.
Read the article →Why proactive proxy rotation makes you more detectable, not less, and what to do instead — block-only rotation, per-profile pairing, cooldown timers, and the rules that extract maximum throughput from a small proxy pool.
Read the article →How to build scrapers that detect failure precisely (block vs. error vs. transient), recover automatically, and produce a clean audit trail — without retry storms, silent corruption, or operator intervention at 2am.
Read the article →How to build a production price-monitoring system in Python — schema design, change detection, alert routing, scaling across thousands of SKUs, and the patterns that distinguish a "tracking spreadsheet" from real competitive intelligence.
Read the article →When to scrape a JavaScript site by intercepting its underlying API calls vs. driving a real browser. The decision framework, the tradeoffs, and how to figure out which approach a target supports in under ten minutes.
Read the article →How to build a production AI extraction pipeline that turns unstructured pages and PDFs into clean, validated, queryable records — covering the scraping front-end, document parsing, LLM-based field extraction, schema validation, and cost discipline.
Read the article →How to take a manual back-office process — spreadsheets, copy-paste, ad-hoc emails — and turn it into a real automation system with queues, workers, an operator dashboard, audit logs, and an exception queue. The patterns that matter and the order to build them in.
Read the article →If an article here lines up with what you're working on, the underlying architecture probably maps directly to something I've already shipped. Reach out and let's talk.
