https://p.agnihotry.com/tags/ai/ Recent content in Ai on Parikshit Agnihotry Hugo en Mon, 09 Feb 2026 00:00:00 -0800 https://p.agnihotry.com/post/designing_a_cost_efficient_agentic_system/ Mon, 09 Feb 2026 00:00:00 -0800 https://p.agnihotry.com/post/designing_a_cost_efficient_agentic_system/ <link rel="stylesheet" href="https://p.agnihotry.com/css/hugo-easy-gallery.css" /> <div class="box" > <figure itemprop="associatedMedia" itemscope itemtype="http://schema.org/ImageObject"> <div class="img"> <img itemprop="thumbnail" src="https://res.cloudinary.com/pagnihotry/image/upload/v1770602020/agentic_systems_main_allsrp.png" alt="https://res.cloudinary.com/pagnihotry/image/upload/v1770602020/agentic_systems_main_allsrp.png"/> </div> <a href="https://res.cloudinary.com/pagnihotry/image/upload/v1770602020/agentic_systems_main_allsrp.png" itemprop="contentUrl"></a> </figure> </div> <p>Extracting deals from emails sounds straightforward until you attempt it at scale. It quickly becomes a deceptively hard NLP problem: unstructured text, image-only creatives, ambiguous marketing language, and no standardized format.</p> <p>Solving it required multiple architectural shifts. What started as prompt-heavy extraction ultimately evolved into a cost-efficient agent-based system.</p> <p>I designed this production-grade pipeline for <a href="http://www.offerivo.com">Offerivo</a> to reliably identify deals, coupons, and expiration dates while keeping inference costs under control.</p>