Bryen Param

Passionate about AI, ML and Technology. Exploring real-world solutions, sharing insights and making sense of complex ideas. #AI #TechInsights #LearningJourney

[Bryen Param @bryen_param]

Hi everyone! I'm Bryen. I just created this account to document my journey in the rapidly evolving world of AI, ML and new tech. #AI #MachineLearning

[Bryen Param @bryen_param]

@satvikputi Nicee 🔥🔥🔥

[Claude @claudeai]

Introducing Claude Opus 4.7, our most capable Opus model yet.

It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back.

You can hand off your hardest work with less supervision.

[Bryen Param @bryen_param]

🎥 Full demo (dashboard view): https://t.co/QvQ3aQEIhd

🚁 Outside view of the drone: https://t.co/xTxrqUwHuZ

💻 Code: https://t.co/sz58r6xxat

Built for the @GoogleDeepMind Gemini Live Agent Challenge. #GeminiLiveAgentChallenge

[Bryen Param @bryen_param]

I always loved how Tony Stark talked to JARVIS ⚡️
So I built something close.
Voice-controlled drone powered by Gemini Live API: it sees, thinks, and responds in real time.
🎥 Demo clip (full video in the replies 👇)
#GeminiLiveAgentChallenge #GoogleAI #Drone #AI

[Bryen Param @bryen_param]

@satvikputi Congrats Satvik ! 🎉

[Thariq @trq212]

We just added /btw to Claude Code!

Use it to have side chain conversations while Claude is working.

[Bryen Param @bryen_param]

brew upgrade claude-code 🥳

[Claude @claudeai]

New in Claude Code: Remote Control.

Kick off a task in your terminal and pick it up from your phone while you take a walk or join a meeting.

Claude keeps running on your machine, and you can control the session from the Claude app or https://t.co/er6Blrr63e

[Bryen Param @bryen_param]

Spent way too long on noisy meditation audio.

Turned out the end of each TTS chunk generates artifacts. Two fixes:

1. Strip the trailing period before sending to ElevenLabs
2. Trim 300ms off the end of each chunk with pydub

Clean silence between chunks. Problem gone.

[Bryen Param @bryen_param]

@safoanTouil @elevenlabs Need to make good use of those credits 👀💪

[Bryen Param @bryen_param]

Building a meditation app with @elevenlabs TTS?

Don't rely on punctuation for long pauses. Use inline SSML break tags:

<break time="2.5s" />

ElevenLabs supports up to 3.0s per tag. Game changer for meditation pacing.

[Bryen Param @bryen_param]

@satvikputi Happy Birthday Satvik ! 🎉🥳

[Bryen Param @bryen_param]

Watching "Her" while building a voice AI project hits different.

I'm using @elevenlabs to generate guided meditations. Still early, but the quality already surprised me.

Maybe Samantha isn't that far away 👀

[Bryen Param @bryen_param]

Just watched 'Her' for the first time in 2026.
Inspiring to see how voice AI could make the world a better place but also a little disturbing how humans connect with it.
Funny how it doesn't feel that far away anymore 😅

[Bryen Param @bryen_param]

@satvikputi @safoanTouil @elevenlabs @SarvamAI Yes! I actually tested Tamil scripts with @elevenlabs, it sounded amazing!
I’ll look into @SarvamAI, thanks for the tip 😉

[Bryen Param @bryen_param]

2 months ago, my teammates and I won the ElevenLabs Paris Hackathon 🏆
So I have a bunch of credits to burn and I'm exploring AI voice use cases.
What would YOU build with it?

[Bryen Param @bryen_param]

Previous demo: "Inspect the bicycle's rear cassette." The AI had to understand bike anatomy, find the rear wheel, locate the cassette.
This one: compare sizes, choose, fly.
Same system. Different intelligence.
Code: https://t.co/PFQhVCmh0i
#Gemini3 #GoogleDeepMind

[Bryen Param @bryen_param]

I can't fly a drone. So I built one that figures it out.
"Find the smallest box on the table." It scans the room. Compares every box. Picks the smallest. Flies to it. No joystick.
A $70 drone + Gemini 3 + two weeks. 🎥👇

[Bryen Param @bryen_param]

@satvikputi Thanks bro! 🙌

[Bryen Param @bryen_param]

I typed:
“Inspect the blue bicycle’s rear cassette.”

The drone figured it out.

No coordinates.
No joystick.
No predefined path.

Gemini 3 analyzes video frames in a perception loop, updates its understanding, and flies itself.

Language → Reasoning → Action

27-sec demo 👇

[Bryen Param @bryen_param]

Built for the @GoogleDeepMind Gemini 3 Hackathon

Demo & project page: https://t.co/r8wgCsil91
Code: https://t.co/PFQhVCmh0i

#Gemini3 #AutonomousAgents #Robotics

[Bryen Param @bryen_param]

How it works:

• Flash → iterative perception over video frames
• Pro → spatial reasoning + inspection report
• Structured output → deterministic drone commands

Frame → Reason → Update → Act → Repeat

An embodied multimodal agent