LLM Models

Large language models for traders: Claude, GPT, Gemini and open models compared, with benchmarks, AI agents and practical guides.

Watercolour of an old Indian mill turned into an AI factory with server racks and a conveyor of glowing slips

Sarvam Token Factory Explained: Token Factory vs Neocloud vs Inference Provider

Token factory explained in simple words: what Sarvam Token Factory is, how neoclouds and inference providers differ, with speed and cost...
Rajandran R
3 min read
Watercolour railway junction where many tracks meet, a metaphor for graph engineering of AI agents

Graph Engineering Explained: How AI Agents Work Together

Graph engineering explained simply: how AI moved from prompt to context, agentic, harness and loop engineering, with trading examples and diagrams.
Rajandran R
11 min read
Watercolor study desk with laptop, brass scale and stopwatch under floating gauges, illustrating LLM metrics

LLM Metrics Explained: Claude Fable 5.2 Benchmarks for Beginners

LLM metrics explained for beginners with Claude Fable 5.2: what each benchmark tests, token price, speed and safety, plus Arena and...
Rajandran R
6 min read
Grid of active and faded stock tiles above a timeline: building survivorship-bias-free NIFTY 50 backtests

How to Build Survivorship-Bias-Free NIFTY 50 Historical Data Using Grokbot

Learn how to build survivorship-bias-free NIFTY 50 backtests using historical constituents, official index weights, effective dates, and NSE Indices data. See...
Rajandran R
5 min read
Chat bubble turning into a trading strategy blueprint about prompting AI to build strategies

How Experienced Traders Should Actually Prompt Claude to Build Strategies

Write a Claude trading strategy prompt that works: specify rules bar by bar, give context, build in stages and avoid the...
Rajandran R
9 min read
Magnifying glass exposing a glitched chart in a chat bubble to catch AI hallucinations

Why AI Hallucinates, How to Catch It, and What Traders and Investors Must Know

AI hallucinations cost traders real capital through fake citations, wrong numbers, and subtly broken backtests. Here is how to catch them...
Rajandran R
8 min read
Two AI chips racing, an unlocked chip closing in on a locked one: open source AI catching up

Open Source AI is Catching Up Fast. Gemma 4 Just Proved It.

The gap between proprietary AI models and open source alternatives has been shrinking for years. But with the release of Google...
Rajandran R
11 min read
Relative rotation graph with coloured quadrants and rotating dot trails around a chip: building an RRG clone

Building an RRG Clone: A Real-World Test of Claude Sonnet 4.5’s Coding Capabilities

Free RRG indicator for TradingView in Pine Script v6, built with Claude Sonnet 4.5: how to use it, read the quadrants,...
Rajandran R
18 min read
Robot lifting a server rack from an old stack into the cloud beside a piggy bank

How I Migrated 5 Years of Website Infrastructure Using AI (and Saved 50% on Hosting)

For five years, MarketCalls ran on the same server setup. Like an old car that still runs but makes concerning noises,...
Rajandran R
2 min read
Split scene of instant code sparkles versus a steady staircase of skill bricks

Agentic Coding and Dopamine Hits: Chasing Quick Wins vs. Building Lasting Skills

Agentic coding feels like magic. One prompt, and whole blocks of code appear on your screen. The dopamine hit is real,...
Rajandran R
1 min read
Neural chip with tuning dials outputting sentiment faces into a chart

Fine-Tuning Gemma 3 270M for Financial Sentiment Analysis using UnSloth

Five days ago, I published an article explaining LoRA and QLoRA as game-changers for traders wanting to fine-tune AI models. The...
Rajandran R
7 min read
Neural network brain with small tuning adapters beside a trading chart

LoRA & QLoRA Explained: How Traders Can Fine-Tune AI Like They Fine-Tune a Strategy

Most traders have heard about AI by now. But let’s be honest , most of it feels too complicated or too...
Rajandran R
1 min read