Ahmad Dogar

404 bio not found

Joined on Jul 9, 2024

Ahmad Dogar

Nov 13

Fixing the Vercel AI SDK Bug That Dropped Gemini File Search Citations in streamText

#opensource #ai #typescript #gemini

9 min read

Want to connect with Ahmad Dogar?

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

Create Account

Already have an account? Sign in

Ahmad Dogar

Nov 11

Grounding AI-written marketing posts in real weather data, without letting the model make things up

#llm #ai #typescript #sqlite

7 min read

Ahmad Dogar

Nov 10

How to evaluate text-to-speech voices before you ship them

#tts #webaudio #javascript #webdev

7 min read

Ahmad Dogar

Nov 8

Atmee V2 vs Seedance: How Causeur and a Real-Time Avatar Model Render in Seconds

#ai #machinelearning #webdev #video

16 min read

Ahmad Dogar

Nov 7

No-login magic links for field crews: hashed, scoped, revocable, and usable offline

#security #webdev #postgres #typescript

7 min read

Ahmad Dogar

Nov 6

SEO for a React SPA without Next.js, and without giving up the 3D homepage

#seo #react #vite #webdev

6 min read

Ahmad Dogar

Nov 5

I built a Claude Code skill that makes video ads without a video model

#ai #claudecode #video #ffmpeg

21 min read

Ahmad Dogar

Nov 5

Agent instructions that actually change behaviour

#ai #claudecode #productivity #devtools

7 min read

Ahmad Dogar

Nov 4

Weight-Space Programming: how GPT-6 Astra edited Qwen's weights and took MATH-500 from 7.9% to 48.1%

#ai #llm #machinelearning #interpretability

22 min read

Ahmad Dogar

Nov 3

How to check an ML benchmark for train/test copies

#machinelearning #python #datascience #testing

11 min read

Ahmad Dogar

Nov 2

GLiNER2 vs an LLM: I tested both on 640 labelled examples

#ai #llm #python #machinelearning

5 min read

Ahmad Dogar

Nov 1

LLMs vs Jev vs CDMs: three ways a model can answer

#ai #llm #machinelearning #performance

13 min read

Ahmad Dogar

Oct 31

Agent loops for product features: budgets, errors, stops and checkpoints

#ai #llm #agents #typescript

8 min read

Ahmad Dogar

Oct 30

Cutting LLM inference cost and latency without making the feature worse

#ai #llm #performance #typescript

8 min read

Ahmad Dogar

Oct 28

Top 10 skills a software engineer needs in 2027

#ai #career #programming #mcp

10 min read

Ahmad Dogar

Oct 27

Caveman mode for Claude Code: fewer output tokens, same facts

#claudecode #ai #llm #productivity

8 min read

Ahmad Dogar

Oct 26

Harness engineering: the skill that took over from prompt engineering

#ai #llm #devtools #productivity

14 min read

Ahmad Dogar

Oct 26

Setting up REA for agent-driven reverse engineering in Claude Code

#reverseengineering #mcp #security #ai

8 min read

Ahmad Dogar

Oct 25

Structured outputs that do not break: schemas, Zod and a repair retry

#ai #typescript #llm #webdev

8 min read

Ahmad Dogar

Oct 25

Evals before prompts: a small eval harness for an LLM feature

#ai #llm #evals #testing

9 min read

Ahmad Dogar

Oct 25

Make coding agents prove their work

#ai #claude #testing #devops

8 min read

DEV Community

Ahmad Dogar

Badges

2 Week Community Wellness Streak

1 Week Community Wellness Streak

Writing Debut

Two Year Club

One Year Club

Fixing the Vercel AI SDK Bug That Dropped Gemini File Search Citations in streamText

Want to connect with Ahmad Dogar?

Grounding AI-written marketing posts in real weather data, without letting the model make things up

How to evaluate text-to-speech voices before you ship them

Atmee V2 vs Seedance: How Causeur and a Real-Time Avatar Model Render in Seconds

No-login magic links for field crews: hashed, scoped, revocable, and usable offline

SEO for a React SPA without Next.js, and without giving up the 3D homepage

I built a Claude Code skill that makes video ads without a video model

Agent instructions that actually change behaviour

Weight-Space Programming: how GPT-6 Astra edited Qwen's weights and took MATH-500 from 7.9% to 48.1%

How to check an ML benchmark for train/test copies

GLiNER2 vs an LLM: I tested both on 640 labelled examples

LLMs vs Jev vs CDMs: three ways a model can answer

Agent loops for product features: budgets, errors, stops and checkpoints

Cutting LLM inference cost and latency without making the feature worse

Top 10 skills a software engineer needs in 2027

Caveman mode for Claude Code: fewer output tokens, same facts

Harness engineering: the skill that took over from prompt engineering

Setting up REA for agent-driven reverse engineering in Claude Code

Structured outputs that do not break: schemas, Zod and a repair retry

Evals before prompts: a small eval harness for an LLM feature

Make coding agents prove their work