Dmytro Dzhulgakov ✓

@dzhulgakov

Co-founder and CTO @FireworksAI_HQ. PyTorch core maintainer. Previously FB Ads. Ex-Pro Competitive Programmer

7 374Followers

817Following

473Posts total

154.6KViews on collected posts

Against accounts of the same size

6 posts from the last 90 days, next to the under 10K follower range. below its peers on both reach and engagement.

Median views7 671this account5 664median for under 10K

Reach, %104.03%this account385.57%median for under 10K

Engagement, %0.72%this account1.10%median for under 10K

Metric	This account	Median for under 10K	Ratio
Median views per post	7 671	5 664	1.35×
Reach (views ÷ followers)	104.03%	3.9× audience	0.27×
Engagement rate	0.72%	1.10%	0.65×

Others in this range → Compare with another account → How these benchmarks are built →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

24.7K29 Nov

21.9K28 Aug

87.3K29 Aug

4.8K

4.2K

10.6K

1.1K

Last 7 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.09%29 Nov

1.57%28 Aug

0.58%29 Aug

0.80%

0.79%

0.34%

0.64%

Reactions — likes, reposts, replies and quotes — divided by views. Median for under 10K accounts is 1.10%.

What the audience does

Likes81.3%879 in total

Reposts3.9%42 in total

Replies4.8%52 in total

Quotes1.3%14 in total

Bookmarks8.7%94 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.

Latest posts

@dzhulgakov Wait, so just to be clear, what you're saying is the https://t.co/6GSKi2RNuZ API was what was wrong? And the longer thinking you were experiencing was actually correct? Also, wow, that's a huge difference between the LLM and SGLang. I would not have guessed that. Inte 1.1K views · 7 likes · 0 reposts · 0 replies 29 Aug 2026 Huge thanks to @Zai_org , @vllm_project and @sgl_project for partnering with us on this investigation Try GLM-5.3-Flash on Fireworks, and let us know if you notice any issues. We also launched GLM-5.3 (Flash’s bigger cousin) same day it was released on Friday. All smooth this 10.6K views · 29 likes · 3 reposts · 2 replies 29 Aug 2026 Along the way, we found a few peculiar behaviors. EvalScope prompt included explicit “think step by step” instructions. It caused Flash to significantly overthink (2-3x more tokens). It was still a fair comparison as we used identical eval harness across endpoints AIME and GPQA 4.2K views · 31 likes · 0 reposts · 1 replies 29 Aug 2026 We believe delaying the launch to investigate quality was the right call. Quality trumps hype. High bar for any launch, gated by public & internal benchmarks and functional tests, sharing fixes with the community - that’s been our consistent focus over the years: DeepSeek V 4.8K views · 35 likes · 1 reposts · 2 replies 29 Aug 2026 GLM-5.3-Flash is live on Fireworks on day… 2 Why? Because we take quality very seriously. We found a benchmark discrepancy we couldn’t explain, so we delayed the launch to investigate. Day 0 (Wed): we saw 2x longer thinking on reasoning-heavy benchmarks (AIME & GPQA) for op 87.3K views · 450 likes · 21 reposts · 28 replies 29 Aug 2026 GLM 5.3 is now live on Fireworks, day zero: → 50% improvement over GLM 5.2 on https://t.co/T8ecPJPEiq’s Code Bench → State-of-the-art open-weight performance on cybersecurity tasks → Excels at complex coding and long-horizon tasks → US-Hosted Serverless on Fireworks Try it http 21.9K views · 307 likes · 16 reposts · 18 replies 28 Aug 2026 Quality from day 1 is something we focus a lot on at @FireworksAI_HQ: including model numerics, tool calling (we identified and fixed bugs even in OpenAI reference code) and reasoning effort control (below) 24.7K views · 20 likes · 1 reposts · 1 replies 29 Nov 2025