At Devlints, I develop backend features for SnagMerch, a product-placement marketplace and OTT/CTV advertising platform. I built an AI-assisted film-scanning pipeline that turns Google Cloud Video Intelligence results into timestamped brand appearances and bounding-box timelines.
I reworked video processing from a blocking job into a quota-aware Celery workflow, with dedicated queues, rate limits, and retries. I also designed a hardened ad-serving API with signed tokens, partner allowlists, and request limits.
For WhatsAppTranscriptor, a Django SaaS for converting WhatsApp chats to PDFs, I designed a seven-app service covering uploads, asynchronous processing, payments, and PDF generation. I made failed jobs recoverable with a resumable state machine and idempotent payment and pipeline handling.
I improved large-export parsing performance and reduced account-deletion time by moving work to a background worker. I also delivered localized PDFs and email in 49 languages and deployed the service as a Docker Compose stack on AWS EC2.
