Chief Product Officer of Air Platform at JetBrains - AI and Agentic Systems
Amsterdam, The Netherlands
Joined May 2009
- Tweets678
- Following729
- Followers440
- Likes506
Air Platform unlocks what I’m most excited about next: the AI Software Factory.
Faster individual developers don’t automatically make a faster organization—and unchecked "dark factories" only produce AI slop.
We’re rethinking software development methodology itself. More on the AI Software Factory soon
We’re bringing all of JetBrains’ AI and agentic experiences together into Air Platform—one cohesive, open system across:
All surfaces: IDEs, web, mobile, Slack, voice, CLI & CI/CD
• All harnesses: Junie, Claude, Codex, Pi & other agents (local + cloud)
• The enterprise & context stack: routing, shared memory, governance & scale
🧵👇
Excited to share that I’ve been promoted to Chief Product Officer of Air Platform at @jetbrains
The future of AI-native software engineering isn’t just a single coding agent—it’s a unified system where developers, teams, and agents work together seamlessly.
Here is what we’re building next 🧵👇
Nick Frolov retweeted
Junie Local now runs on @NVIDIARTXSpark from day one.
Get a full-fledged coding agent on your PC. It reads your code, edits files, and runs tests. There are no tokens or quotas, and no code leaves the machine.
Windows support is stable, too. All you need is a 24 GB Nvidia RTX card.
Try it for free: junie.jetbrains.com/local
ALT Image with a black background features text: "Your RTX Spark just got an agent. Junie Local runs on it from day one." The text is in white and green. The top corners display the logos for Junie and JetBrains. The bottom shows a green command line: "> /local".
Nick Frolov retweeted
Your GPU. Your code. No token bill.
Junie Local runs on NVIDIA RTX Spark from day one. Windows support is stable, too.
Reads code. Edits files. Runs tests. All locally.
Try it free → junie.jetbrains.com/blog/jun…
🤖 Made with AI
In our testing on the RTX Spark, Junie Local reached generation speeds of up to 55 tokens per second for Qwen 27b model
Junie now runs local LLM models on Nvidia RTX Spark laptops 🔥
RTX Spark platform pairs a Blackwell RTX GPU with ARM-based Grace CPU allowing to use unified memory on the laptop.
Don’t have RTX Spark powered laptops yet? We also support Apple M5 and NVIDIA RTX 3090, 4090 and 5090.
Install @junie_ai and run /local, and you have a full coding agent on your RTX Spark. It reads your code, edits files, and runs your tests. No token charges, no quotas, and your code never leaves the machine.
blog.jetbrains.com/junie/202…
Nick Frolov retweeted
We connected IntelliJ IDEA’s code intelligence to Claude Code and Codex, then tested the setup with Netflix.
The first public results are out.
Typical efficiency gains: 20–30%.
Best code-navigation case: ~65%.
These results come from two real codebases. The efficiency metric averages token savings and time savings.
Correctness stayed stable or improved.
Here’s what we actually built.
The code engine behind IntelliJ IDEA ran headlessly, giving agents access to its capabilities through a CLI backed by a daemon.
LSP alone wasn’t sufficient.
Agents needed a convenient way to invoke the tools. The backend also needed to stay running and keep the project index warm between calls.
The idea is straightforward.
An IDE has already mapped the structure of your code and its dependencies — including what’s inside compiled libraries.
An agent can use that knowledge instead of spending time and tokens reconstructing it.
But the most interesting finding was which models benefited most.
The strongest models saw the biggest gains.
Especially at higher reasoning effort.
In these experiments, more capable models were also better at taking advantage of code intelligence.
My takeaway: better models and better tools aren’t competing bets.
You don’t have to choose between investing in model intelligence and giving agents better access to your codebase.
The two can reinforce each other.
Huge thanks to @pbakker and the Netflix team for working with us on this!
Nick Frolov retweeted
We want you to know that Air lets you run agents in the cloud, too.
Initially, we recorded this demo on a small project, but it felt like cheating. So we re-shot it on the real IntelliJ IDEA repository, where the team works every day.
When market categories get redefined within a year products restructuring is inevitable.
Hey, at least It’s never boring 😜
@kskrygan presents JetBrains Air at @WorldSummitAI
JetBrains Air is new system for building software with agents
Nick Frolov retweeted
Automated debugging by AI agents in Air with IntelliJ IDEA. Such a quality-of-life improvement!
Big Day for @jetbrains IDEs and for @junie_ai .
A completely new agentic experience is open today in our IDEs - we call it JetBrains Air. And the best part - Junie provides free baseline intelligence tier - Junie Lite.
In JetBrains Air you have a professional control plane over your agents (plug-in any agent you already love), agents work in parallel, you can link sessions with each other, delegate inline comments from code or md files to agents and many more.
Read the details here: blog.jetbrains.com/ai/2026/1…
Nick Frolov retweeted
We've reimagined JetBrains IDEs to become the ultimate agentic workbench.
Air is an open hub for agents inside your IDE:
-Any agents to run (Claude, Codex & many more)
-Agents are more efficient with IDE intelligence
-Free Junie Lite agent bundled
-Cloud runs via Air Teams
⬇️
@junie_ai team is attending Gemini 4 Argon evaluation workshop. The latest model from @GoogleDeepMind
The Evolution of Voice as UI:
1. Hardcoded Intents: TTS and STT have been largely solved for years, but early assistants like Siri and Alexa relied on rigid, prescripted intent mapping.
2. Raw Transcription: Voice input just captured verbatim speech filler words, false starts, awkward pauses, and all.
3. LLM Cleaners: Tools like Whisper Flow arrived, using LLMs as a harness to turn messy human speech into structured, coherent text prompts.
4. Multimodal Local Agents: Systems like Codex integrated voice with vision and direct on-device action.
5. Remote Agent Control: With Meta Muse and OpenAI's latest agentic workflows, the compute is entirely remote. Hardware like Meta RayBans or your phone becomes pure audio I/O, an ambient steering wheel for an agent running a full computer in the cloud.
