firefrog

I am a developer who is inspired by the beauty underlies the logic of algorithms. Also, I am a fan of AI. I do believe that one day, human can program human.

Viet Nam Joined on Jan 5, 2024 minh.leduc.0210@gmail.com https://minlee0210.github.io/

firefrog

Sep 27

Jev's Boring Use Cases Are the Ones That Actually Work

#ai #jev #typesafeai

4 min read

Want to connect with firefrog?

Create an account to connect with firefrog. You can also sign in below to proceed if you already have an account.

Create Account

Already have an account? Sign in

firefrog

Sep 26

I Reproduced RLCD to Find Out What Its "Calibrated RL" Actually Does

5 min read

firefrog

Sep 26

I Took Laya's "Calibrated RL" Apart. It's Cross-Entropy With Extra Steps.

#ai #aiagents #calibration #jev

6 min read

firefrog

Sep 26

Inside RLCD: The Estimator, the Proof, and the 30 Runs That Changed My Mind

#ai #jev #laya #typesafe

8 min read

firefrog

Sep 26

Pi‑warden: Using Jev to Block Destructive Commands in 48 Hours

#agenticai #ai #aiagents #jev

4 min read

firefrog

Sep 25

Understanding RLCD: Calibration Techniques for Decision‑Only LLMs

#ai #aiagents #jev #typesafeai

4 min read

firefrog

Sep 24

Jev, Explained: The AI That Refuses to Write a Word

#ai #jev #llm

4 min read

firefrog

Sep 23

SmolVLM2 vs Qwen2.5-VL: Real-Time RTSP Edge Video Summarization Under 8GB VRAM

#ai #edgeai

10 min read

firefrog

Sep 22

From Codebase to Cognitive Skill: Inside BAAI's Repo-to-Skill Framework

#agentharness #agenticai #ai #contextengineering

12 min read

firefrog

Sep 17

Dynamic Reasoning Budgets: Halving Inference Costs on Test-Time Compute Models

#ai

6 min read

firefrog

Sep 16

The Harness is the Moat: Building a Deterministic Agent Runtime with Context Pruning

#ai

7 min read

firefrog

Sep 10

I Tested 3 LLMs on Vietnamese Idioms. They Got It Backwards.

7 min read

firefrog

Sep 9

AI Reads Vietnamese Slang as Angry. 3 Models Don't.

7 min read

firefrog

Sep 8

The Real Vietnamese LLM Tokenizer Cost (It's Not 4.5x)

9 min read

firefrog

Sep 6

A Simple Look at TokPress: A Compressor That Uses an LLM's Tokenizer

3 min read

firefrog

Sep 5

Agent harness or agent framework?

6 min read

firefrog

Sep 2

Five ways to invalidate your prompt cache

9 min read

firefrog

Sep 1

Fun Project: I Built a Compressor That Thinks in Tokens

10 min read

firefrog

Aug 15

How do you deduplicate a stream you don't control?

#architecture #backend #data #software

6 min read

firefrog

Jul 19

A Hands-On Guide to kalbee: Your First Kalman Filter (and Beyond)

#ai #computervision #opensource #tutorial

5 min read

firefrog

Jul 14

Meet kalbee: State Estimation and Target Tracking, Minus the Pain

2 min read

firefrog

Jul 11

wavio

#opensource #python #learning #startup

1 min read

firefrog

Jul 2

Enigmar

#algorithms #performance #python #rust

3 min read

firefrog

Sep 2 '24

How to build an autonomous news Generator with AI using Fluvio

#challenge #ai #pubsub #webdev

10 min read

firefrog

Aug 28 '24

How to build an event-driven architecture with Fluvio

#webdev #eventdriven #python #pubsub

3 min read

firefrog

Jul 31 '24

Boost Your RAG Performance with Tavily Search API

#ai #nlp #api #python

2 min read

DEV Community

firefrog

Badges

Two Year Club

One Year Club

Writing Debut

Jev's Boring Use Cases Are the Ones That Actually Work

Want to connect with firefrog?

I Reproduced RLCD to Find Out What Its "Calibrated RL" Actually Does

I Took Laya's "Calibrated RL" Apart. It's Cross-Entropy With Extra Steps.

Inside RLCD: The Estimator, the Proof, and the 30 Runs That Changed My Mind

Pi‑warden: Using Jev to Block Destructive Commands in 48 Hours

Understanding RLCD: Calibration Techniques for Decision‑Only LLMs

Jev, Explained: The AI That Refuses to Write a Word

SmolVLM2 vs Qwen2.5-VL: Real-Time RTSP Edge Video Summarization Under 8GB VRAM

From Codebase to Cognitive Skill: Inside BAAI's Repo-to-Skill Framework

Dynamic Reasoning Budgets: Halving Inference Costs on Test-Time Compute Models

The Harness is the Moat: Building a Deterministic Agent Runtime with Context Pruning

I Tested 3 LLMs on Vietnamese Idioms. They Got It Backwards.

AI Reads Vietnamese Slang as Angry. 3 Models Don't.

The Real Vietnamese LLM Tokenizer Cost (It's Not 4.5x)

A Simple Look at TokPress: A Compressor That Uses an LLM's Tokenizer

Agent harness or agent framework?

Five ways to invalidate your prompt cache

Fun Project: I Built a Compressor That Thinks in Tokens

How do you deduplicate a stream you don't control?

A Hands-On Guide to kalbee: Your First Kalman Filter (and Beyond)

Meet kalbee: State Estimation and Target Tracking, Minus the Pain

wavio

Enigmar

How to build an autonomous news Generator with AI using Fluvio

How to build an event-driven architecture with Fluvio

Boost Your RAG Performance with Tavily Search API