23 | Devtools ❤️ | Engineer @ona_hq | Maintainer @OpenFGA, github1s 💚 | Building agent infra & tools | Author of agents.siddhantkhare.com | opinions are mine
Joined July 2018
- Tweets 9,534
- Following 1,517
- Followers 5,843
- Likes 21,311
1,483 Photos and videos
Pinned Tweet
I wrote a book for engineering leaders.
The Agentic Engineering Guide. 10 parts, 33 chapters. Context engineering, authorization, cost control, sandboxing, evaluation, governance, adoption playbooks.
Not a prompt tutorial. The decisions you face when shipping agents to production this quarter.
Early version. Open source. Rough edges included. Spot a mistake, open a PR.
agents.siddhantkhare.com
5
39
328
30,600
Jun 11
It’s been a great journey from joining as an Intern at the seed stage to the acquisition!!!
More to come in my blog post, all my learnings and memories ❤️
Jun 11
We’ve reached an agreement to acquire @ona_hq.
Its secure cloud execution technology will help Codex take on longer-running work, even when laptops are closed, and help more organizations deploy agents securely in production.
After closing, Ona will join OpenAI’s Codex team. openai.com/index/openai-to-a…
23
2
167
19,754
Blogpost is out now
I started at Gitpod at 19.
Three-month internship. I spent my first week trying random things. I just kept going.
Four years later, @ona_hq is joining @OpenAI
I wrote about the path b/w those two moments.
Also open to what comes next. Agent infra, devtools, developer platforms.
siddhantkhare.com/writing/fr…
79
Jun 11
🚀❤️
Jun 11
We’ve reached an agreement to acquire @ona_hq.
Its secure cloud execution technology will help Codex take on longer-running work, even when laptops are closed, and help more organizations deploy agents securely in production.
After closing, Ona will join OpenAI’s Codex team. openai.com/index/openai-to-a…
15
134
27,490
Blogpost is out now -
I started at Gitpod at 19.
Three-month internship. I spent my first week trying random things. I just kept going.
Four years later, @ona_hq is joining @OpenAI
I wrote about the path b/w those two moments.
Also open to what comes next. Agent infra, devtools, developer platforms.
siddhantkhare.com/writing/fr…
42
I started at Gitpod at 19.
Three-month internship. I spent my first week trying random things. I just kept going.
Four years later, @ona_hq is joining @OpenAI
I wrote about the path b/w those two moments.
Also open to what comes next. Agent infra, devtools, developer platforms.
siddhantkhare.com/writing/fr…
11
4
60
3,048
cried multiple times while writing this, tried to wrote my heart out 🥹❤️
4
227
Jun 11
Jun 11
It’s been a great journey from joining as an Intern at the seed stage to the acquisition!!!
More to come in my blog post, all my learnings and memories ❤️
2
1,830
Siddhant Khare retweeted
Jun 10
Added Codex & Gemini CLI support for Agent trace!!!
1
1
7
844
Jun 10
Added Codex & Gemini CLI support for Agent trace!!!
1
1
7
844
Jun 10
4
282
Jun 10
One thing I observed is that when you build something like this, which is easy to demo and to explain, it gets more traction & love comparable to the AI Infra-related stuff that I build
I saw the Gemma 4 announcement two days ago, found some time tonight. Built a small CLI around it.
Gemma 4 12B can describe images, and it's just an 8GB model you can self-host, which means you can wire it into small utility tools without paying for a vision API.
I used it to generate alt text & blog outlines from whiteboard photos
Works well enough to be useful :D
1
1
5
375
found this pretty interesting paper on context de-duplication. It does byte-exact deduplication for LLM inference. It's a different approach (byte-level hashing vs. my Distill's semantic clustering MMR reranking), but covers similar ground
diff: Merlin is byte-exact & lossless. Distill is semantic & lossy-by-design
1
3
384
read paper: arxiv.org/abs/2605.09990
2
260
I am really loving the powerful local model Gemma4
this a quick utitlty Golang based CLI tool to do the lipsync check in videos
1
14
1,035
Siddhant Khare retweeted
I saw the Gemma 4 announcement two days ago, found some time tonight. Built a small CLI around it.
Gemma 4 12B can describe images, and it's just an 8GB model you can self-host, which means you can wire it into small utility tools without paying for a vision API.
I used it to generate alt text & blog outlines from whiteboard photos
Works well enough to be useful :D
3
6
39
8,592
I saw the Gemma 4 announcement two days ago, found some time tonight. Built a small CLI around it.
Gemma 4 12B can describe images, and it's just an 8GB model you can self-host, which means you can wire it into small utility tools without paying for a vision API.
I used it to generate alt text & blog outlines from whiteboard photos
Works well enough to be useful :D
3
6
39
8,592
Since December, I’ve primarily built 3 open-source products on weekends.
They still don’t look impressive if you only judge them by GitHub stars, but all three have led to VC interest. More importantly, the first two have also started getting meaningful enterprise usage.
Distill was the first one that helped me connect with some good VCs, but after building AgentTrace and AgentAuthz, the conversations became much stronger, and now there are actual enterprise users as well.
the reson I’m writng all this that you don’t always need a huge number of GitHub stars to get interest, users, or investors.
Stars can help, of course. But people who actually need the tool will find it, use it & sometimes even invest in it
2
1
28
935