DEV Community

Devanshu Biswas profile picture

Devanshu Biswas

404 bio not found

Joined Joined on 
AI Agents Explained: the Thought-Action-Observation Loop

AI Agents Explained: the Thought-Action-Observation Loop

1
Comments
1 min read
Fine-tuning vs RAG: Two Ways to Teach an LLM

Fine-tuning vs RAG: Two Ways to Teach an LLM

Comments
1 min read
Temperature and Sampling: the LLM Creativity Dial

Temperature and Sampling: the LLM Creativity Dial

Comments
1 min read
The Context Window: an LLM's Short-Term Memory, Explained

The Context Window: an LLM's Short-Term Memory, Explained

1
Comments
1 min read
Why LLMs Hallucinate, and How to Reduce It

Why LLMs Hallucinate, and How to Reduce It

1
Comments
2 min read
LSTM and GRU: the Gates That Let a Network Remember

LSTM and GRU: the Gates That Let a Network Remember

Comments
1 min read
RNNs From Scratch: a Network With a Memory of the Past

RNNs From Scratch: a Network With a Memory of the Past

Comments
1 min read
Pooling in CNNs: Shrink the Map, Keep What Matters

Pooling in CNNs: Shrink the Map, Keep What Matters

Comments
1 min read
CNNs From Scratch: the Convolution That Sees Images

CNNs From Scratch: the Convolution That Sees Images

Comments
1 min read
Optimizers Explained: SGD vs Momentum vs Adam, Visualized

Optimizers Explained: SGD vs Momentum vs Adam, Visualized

Comments
1 min read
Random Forest From Scratch: Many Weak Trees Beat One Strong Tree

Random Forest From Scratch: Many Weak Trees Beat One Strong Tree

Comments
1 min read
Feature Scaling: Why One Big Number Drowns Your Model

Feature Scaling: Why One Big Number Drowns Your Model

Comments
1 min read
Why Accuracy Lies: Precision, Recall and F1 From Scratch

Why Accuracy Lies: Precision, Recall and F1 From Scratch

Comments
1 min read
Overfitting Made Visible: Train/Test and the U-Shaped Error Curve

Overfitting Made Visible: Train/Test and the U-Shaped Error Curve

Comments
1 min read
Gradient Descent From Scratch: the Engine Under All of ML

Gradient Descent From Scratch: the Engine Under All of ML

Comments
1 min read
I Built a Live Markdown Editor With No Library (~40 Lines)

I Built a Live Markdown Editor With No Library (~40 Lines)

Comments
1 min read
I Built a QR Code Generator That Runs 100% in the Browser

I Built a QR Code Generator That Runs 100% in the Browser

Comments
1 min read
Self-RAG: Let the Model Decide When to Retrieve, Then Grade Itself

Self-RAG: Let the Model Decide When to Retrieve, Then Grade Itself

Comments
1 min read
I Built a Toast Notification System in ~40 Lines (Stack, Auto-Dismiss, Variants)

I Built a Toast Notification System in ~40 Lines (Stack, Auto-Dismiss, Variants)

Comments
1 min read
I Built Pac-Man in ~150 Lines of Vanilla JavaScript

I Built Pac-Man in ~150 Lines of Vanilla JavaScript

1
Comments
1 min read
I'm Building a Production-Grade Spring Boot + React App, One Feature a Day (Day 1)

I'm Building a Production-Grade Spring Boot + React App, One Feature a Day (Day 1)

1
Comments
2 min read
What an LLM Actually Does: Predicting the Next Word, Explained

What an LLM Actually Does: Predicting the Next Word, Explained

Comments
2 min read
Loss Functions: MSE vs MAE vs Cross-Entropy, Visualized

Loss Functions: MSE vs MAE vs Cross-Entropy, Visualized

1
Comments
2 min read
Naive Bayes From Scratch: A Spam Filter Built From Word Counts

Naive Bayes From Scratch: A Spam Filter Built From Word Counts

1
Comments
2 min read
I Built an Image Compressor That Runs 100% in the Browser

I Built an Image Compressor That Runs 100% in the Browser

1
Comments
2 min read
RAG-Fusion: Ask the Question Many Ways, Then Fuse the Ranks (RRF)

RAG-Fusion: Ask the Question Many Ways, Then Fuse the Ranks (RRF)

Comments
2 min read
I Rebuilt the Cmd-K Command Palette in ~60 Lines of JavaScript

I Rebuilt the Cmd-K Command Palette in ~60 Lines of JavaScript

Comments
2 min read
I Built Asteroids in ~120 Lines of Vanilla JavaScript

I Built Asteroids in ~120 Lines of Vanilla JavaScript

2
Comments 1
2 min read
WebGPU: I Simulated 250,000 Particles Entirely on the GPU (in the Browser)

WebGPU: I Simulated 250,000 Particles Entirely on the GPU (in the Browser)

Comments
2 min read
How a Neural Network Actually Learns — Training, in Plain Words

How a Neural Network Actually Learns — Training, in Plain Words

1
Comments
2 min read
The XOR Problem: Why Neural Nets Needed Hidden Layers

The XOR Problem: Why Neural Nets Needed Hidden Layers

Comments
2 min read
Decision Trees From Scratch: Yes/No Questions That Carve Up Data

Decision Trees From Scratch: Yes/No Questions That Carve Up Data

Comments
2 min read
I Built Live Captions in the Browser — No API Key, No Server

I Built Live Captions in the Browser — No API Key, No Server

Comments
2 min read
HyDE: The RAG Trick Where You Answer the Question BEFORE You Search

HyDE: The RAG Trick Where You Answer the Question BEFORE You Search

1
Comments
2 min read
I Rebuilt Notion's Slash Menu (/) in ~70 Lines of JavaScript

I Rebuilt Notion's Slash Menu (/) in ~70 Lines of JavaScript

1
Comments
2 min read
I Built Space Invaders in ~120 Lines of Vanilla JavaScript

I Built Space Invaders in ~120 Lines of Vanilla JavaScript

1
Comments 1
2 min read
I Built a Real Weather App With React Native — Same React, Now on Your Phone

I Built a Real Weather App With React Native — Same React, Now on Your Phone

1
Comments 1
2 min read
I Built a Webcam Sign-Language Reader in the Browser (No Cloud)

I Built a Webcam Sign-Language Reader in the Browser (No Cloud)

Comments
3 min read
RAG Rerank: the Highest-Leverage Upgrade to Your Retrieval Pipeline

RAG Rerank: the Highest-Leverage Upgrade to Your Retrieval Pipeline

2
Comments
3 min read
I Rebuilt Slack's Message Box With Plain HTML, CSS and JS

I Rebuilt Slack's Message Box With Plain HTML, CSS and JS

2
Comments
2 min read
I Built Frogger in 120 Lines of Vanilla JavaScript

I Built Frogger in 120 Lines of Vanilla JavaScript

Comments
2 min read
What a Neural Net Actually Does — the Intuition, No Math

What a Neural Net Actually Does — the Intuition, No Math

Comments
3 min read
Backpropagation From Scratch: How a Neural Network Actually Learns

Backpropagation From Scratch: How a Neural Network Actually Learns

Comments
3 min read
K-Means From Scratch: Finding Groups in Data With No Labels

K-Means From Scratch: Finding Groups in Data With No Labels

1
Comments
3 min read
I Built a Real Phone App With Flutter — One Codebase for iOS + Android

I Built a Real Phone App With Flutter — One Codebase for iOS + Android

Comments
3 min read
I Built Speech-to-Text That Runs 100% On My Machine — No Cloud, No API Key (Whisper)

I Built Speech-to-Text That Runs 100% On My Machine — No Cloud, No API Key (Whisper)

1
Comments
4 min read
I Built a Study-Notes Summarizer in the Browser — No AI API, Just Word-Frequency Scoring

I Built a Study-Notes Summarizer in the Browser — No AI API, Just Word-Frequency Scoring

1
Comments
4 min read
I Built Flappy Bird in 120 Lines of Vanilla JavaScript (No Engine)

I Built Flappy Bird in 120 Lines of Vanilla JavaScript (No Engine)

2
Comments
5 min read
Embeddings Explained Simply: How AI Turns Words Into a Map of Meaning

Embeddings Explained Simply: How AI Turns Words Into a Map of Meaning

Comments
3 min read
ReAct + Tools: How an LLM Actually Uses a Calculator and a Search Engine

ReAct + Tools: How an LLM Actually Uses a Calculator and a Search Engine

1
Comments
5 min read
I Cloned the Discord Chat UI With Pure CSS — Here's How the Layout Works

I Cloned the Discord Chat UI With Pure CSS — Here's How the Layout Works

Comments
3 min read
The Forward Pass, Demystified: How a Neural Network Turns Inputs Into an Answer

The Forward Pass, Demystified: How a Neural Network Turns Inputs Into an Answer

Comments
3 min read
k-Nearest Neighbors From Scratch: the ML Algorithm With No Training Step

k-Nearest Neighbors From Scratch: the ML Algorithm With No Training Step

Comments
4 min read
Tokens: Why ChatGPT Can't Count the R's in 'Strawberry'

Tokens: Why ChatGPT Can't Count the R's in 'Strawberry'

Comments 1
1 min read
Activation Functions: Why a 100-Layer Network Without Them Is Still One Line

Activation Functions: Why a 100-Layer Network Without Them Is Still One Line

Comments
1 min read
Logistic Regression From Scratch: the Same Gradient Descent, Squashed Through a Sigmoid

Logistic Regression From Scratch: the Same Gradient Descent, Squashed Through a Sigmoid

Comments
2 min read
A Resume Bullet Improver in 40 Lines of Regex (No AI Needed)

A Resume Bullet Improver in 40 Lines of Regex (No AI Needed)

Comments
1 min read
Tree of Thoughts: How to Make an LLM Explore Instead of Guess

Tree of Thoughts: How to Make an LLM Explore Instead of Guess

Comments
2 min read
Spotify's Now-Playing Bar Is Just One CSS Grid + a Ticking Width

Spotify's Now-Playing Bar Is Just One CSS Grid + a Ticking Width

Comments
1 min read
I Built Minesweeper in ~50 Lines — the Only Hard Part Is Flood-Fill

I Built Minesweeper in ~50 Lines — the Only Hard Part Is Flood-Fill

Comments
1 min read
loading...