Thomas Ashish Cherian PandaWhoCodes

Slashing LLM Token Costs for Insurance PDF Extraction

Context: A pipeline that fetches emails, triages them via LLM, downloads PDF attachments, extracts insurance policy details via LLM, and saves results. Currently spending ~$0.038/run on Groq Llama 4 Scout.

Current State — Where the Money Goes

Per Run (1 user):

Offloading Processing to On-Demand Workers — Complete Guide

Context: A FastAPI app running a sequential email→triage→PDF→LLM→save pipeline (30-60s per run). Goal: cut always-on compute costs and scale individual bottlenecks independently.

Current Architecture — What We're Working With

User Request → FastAPI (always-on server)

Build2Learn.in — Pivoting in the Age of AI

The Catalyst

"Code is cheap. Show me the talk." — nadh.in/blog/code-is-cheap

Software development as we knew it is changing. LLMs have made building easier — anyone can spin up a clone of anything. So what matters now? Ideas, marketing, go-to-market, and real-world problem solving. This sparked a discussion among Build2Learn's core members on where to take the community next.

Build First, Think Never: The Wisdom We Keep Forgetting

In which humanity repeatedly chooses the mirage of quick profits over the hard work of long-term thinking—and why our relationship with AI is following the same tragic script

The Wisdom of Socrates and the Folly of Progress

In ancient Athens, when Socrates first encountered the revolutionary technology of writing, he was deeply troubled. Not because he opposed innovation, but because he understood something that modern innovators have forgotten: every technology is a Faustian bargain.

	<!DOCTYPE html>
	<html lang="en">
	<head>
	<meta charset="UTF-8">
	<meta name="viewport" content="width=device-width, initial-scale=1.0">
	<title>Things to Do in Chennai — 80+ Activities Guide</title>
	<link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700;800;900&display=swap" rel="stylesheet">
	<script src="https://cdn.tailwindcss.com"></script>
	<script src="https://cdnjs.cloudflare.com/ajax/libs/gsap/3.12.5/gsap.min.js"></script>
	<script src="https://cdnjs.cloudflare.com/ajax/libs/gsap/3.12.5/ScrollTrigger.min.js"></script>

	<!DOCTYPE html>
	<html lang="en" class="scroll-smooth">
	<head>
	<meta charset="UTF-8">
	<meta name="viewport" content="width=device-width, initial-scale=1.0">
	<title>Chennai Explorer — 100+ Things to Do</title>
	<link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700;800;900&display=swap" rel="stylesheet">
	<script src="https://cdn.tailwindcss.com"><\/script>
	<script src="https://cdnjs.cloudflare.com/ajax/libs/gsap/3.12.5/gsap.min.js"><\/script>
	<script src="https://cdnjs.cloudflare.com/ajax/libs/gsap/3.12.5/ScrollTrigger.min.js"><\/script>

	A
	A + Certified
	A-110
	A-122
	A-123
	A-133
	A-frames
	A-GPS
	A/B Testing
	A/R analysis

	To sell only the stocks that are not there in this week's new strategy and keep the stocks that still remain, you can modify the code as follows:

	```python
	import backtrader as bt
	import pandas as pd

	# Define the strategy
	class MyStrategy(bt.Strategy):

	def __init__(self):