DEV Community

xbill profile picture

xbill Subscriber

Master of MCP and herding cats

Location New York Joined Joined on  Email address xbill@glitnir.com Personal website https://g.dev/xbill github website
Community Favorite 2
Summer Bug Smash Winner
DEV Weekend Challenge Completion
2
Frontend Challenge Completion Badge
One Year Club
Cloud Run Badge
Go
June Solstice Game Jam Completion
Azure
GitHub Copilot "Finish-Up-A-Thon" Challenge Completion
Gemma 4 Challenge Completion
Hermes Agent Challenge Completion

Pin Pinned

The unofficial TPU migration guide: Cloud TPU API to Compute Engine
xbill profile

The unofficial TPU migration guide: Cloud TPU API to Compute Engine

#tpu #gcp #vllm #devops
8
Comments 2
17 min read
Three Clouds, Three Native Agents
xbill profile

Three Clouds, Three Native Agents

#cloud #ai #architecture #devops
Comments 2
10 min read
Five Gemma-4 models, one accelerator: what porting E2B 31B to AWS Inferentia2 taught me
xbill profile

Five Gemma-4 models, one accelerator: what porting E2B 31B to AWS Inferentia2 taught me

#machinelearning #aws #ai #python
2
Comments
6 min read
Why did my benchmark stop at N=22? A debugging story in nine bugs

Summer Bug Smash: Smash Stories Submission 🐛🛹
xbill profile

Why did my benchmark stop at N=22? A debugging story in nine bugs

#bugsmash #debugging #ai #python
13
Comments 6
4 min read
Mix and Match: Running Kiro on Azure Cloud Shell
xbill profile

Mix and Match: Running Kiro on Azure Cloud Shell

#cli #cloudshell #azure #aws
1
Comments
6 min read
Write Markdown Once, Publish It Everywhere: dev.to, Medium, AWS Builder Center and LinkedIn
xbill profile

Write Markdown Once, Publish It Everywhere: dev.to, Medium, AWS Builder Center and LinkedIn

#writing #markdown #devtools #ai
5
Comments
10 min read

Want to connect with xbill?

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

Create Account
Already have an account? Sign in
Write Markdown Once, Publish It Everywhere: dev.to, Medium, AWS Builder Center and LinkedIn
xbill profile

Write Markdown Once, Publish It Everywhere: dev.to, Medium, AWS Builder Center and LinkedIn

#writing #markdown #devtools #ai
17
Picked as gem Comments 3
10 min read
Repacked QAT Gemma 4 on One TPU v5e: 12B Serves at 675 Tokens per Second
xbill profile

Repacked QAT Gemma 4 on One TPU v5e: 12B Serves at 675 Tokens per Second

#gemma #googlecloud #machinelearning #llm
8
Comments
9 min read
Gemma 4 QAT on One TPU v5e: What Runs and What Doesn't
xbill profile

Gemma 4 QAT on One TPU v5e: What Runs and What Doesn't

#gemma #googlecloud #machinelearning #llm
6
Comments
9 min read
Gemma 4 at Over 70 Tokens/s on a 2021 Laptop's 4 GB GPU: The Live Demo, Step by Step
xbill profile

Gemma 4 at Over 70 Tokens/s on a 2021 Laptop's 4 GB GPU: The Live Demo, Step by Step

#gemma #llamacpp #cuda #huggingface
9
Comments
12 min read
Gemma 4 on Amazon SageMaker or a VM? The Same Model Server at 1.40x the Price
xbill profile

Gemma 4 on Amazon SageMaker or a VM? The Same Model Server at 1.40x the Price

#aws #sagemaker #gemma #vllm
1
Comments
8 min read
Gemma 4 on Amazon SageMaker: The NVIDIA T4 Decodes at 0.8x of the L4 With the Same Answers
xbill profile

Gemma 4 on Amazon SageMaker: The NVIDIA T4 Decodes at 0.8x of the L4 With the Same Answers

#aws #sagemaker #gemma #vllm
15
Picked as gem Comments 1
9 min read
Gemma 4 on Amazon SageMaker: The NVIDIA T4 Decodes at 0.8x of the L4 With the Same Answers
xbill profile

Gemma 4 on Amazon SageMaker: The NVIDIA T4 Decodes at 0.8x of the L4 With the Same Answers

#aws #sagemaker #gemma #vllm
1
Comments
9 min read
Gemma 4 on a Tesla T4, Part 3: Int4 Embeddings Serve E2B in 2.86 GiB at 2.30x bf16
xbill profile

Gemma 4 on a Tesla T4, Part 3: Int4 Embeddings Serve E2B in 2.86 GiB at 2.30x bf16

#gemma #vllm #cuda #machinelearning
9
Comments
11 min read
Gemma 4 on Amazon SageMaker: 4-Bit Embeddings Decode up to 1.39x Faster on One L4
xbill profile

Gemma 4 on Amazon SageMaker: 4-Bit Embeddings Decode up to 1.39x Faster on One L4

#aws #sagemaker #gemma #vllm
1
Comments
11 min read
Count It or Compute It: When a Tool Returns Rows, the Models That Count Them Right Spend the Tokens

Kaggle Benchmarking Challenge Submission
xbill profile

Count It or Compute It: When a Tool Returns Rows, the Models That Count Them Right Spend the Tokens

#devchallenge #kagglechallenge #ai #machinelearning
15
Comments 7
10 min read
Google's QAT Gemma 4 26B-A4B on One TPU v6e: 15.6x the KV Cache and 1.9x the Throughput of FP8
xbill profile

Google's QAT Gemma 4 26B-A4B on One TPU v6e: 15.6x the KV Cache and 1.9x the Throughput of FP8

#gemma #googlecloud #machinelearning #llm
Comments 1
12 min read
Gemma 4 on Amazon SageMaker: QAT Weights Decode 2.05x Faster Than bf16 on One L4
xbill profile

Gemma 4 on Amazon SageMaker: QAT Weights Decode 2.05x Faster Than bf16 on One L4

#aws #sagemaker #gemma #vllm
7
Comments
9 min read
Gemma 4 on an Amazon SageMaker Endpoint: AWS CLI, NVIDIA L4, and an MCP Server
xbill profile

Gemma 4 on an Amazon SageMaker Endpoint: AWS CLI, NVIDIA L4, and an MCP Server

#aws #sagemaker #gemma #mcp
7
Comments
10 min read
Gemma 4 on Amazon SageMaker: QAT Weights Decode 2.05x Faster Than bf16 on One L4
xbill profile

Gemma 4 on Amazon SageMaker: QAT Weights Decode 2.05x Faster Than bf16 on One L4

#aws #sagemaker #gemma #vllm
Comments
9 min read
Gemma 4 on an Amazon SageMaker Endpoint: AWS CLI, NVIDIA L4, and an MCP Server
xbill profile

Gemma 4 on an Amazon SageMaker Endpoint: AWS CLI, NVIDIA L4, and an MCP Server

#aws #sagemaker #gemma #mcp
Comments
10 min read
Two Iceberg Clients, One Protocol: Where the Time Goes
xbill profile

Two Iceberg Clients, One Protocol: Where the Time Goes

#rust #python #iceberg #performance
8
Comments
9 min read
Two Iceberg Clients, One Protocol: Where the Time Goes
xbill profile

Two Iceberg Clients, One Protocol: Where the Time Goes

#rust #python #iceberg #performance
Comments
9 min read
What One Rust Client Can Reach Across Seven Iceberg Catalogs
xbill profile

What One Rust Client Can Reach Across Seven Iceberg Catalogs

#rust #iceberg #lakehouse #dataengineering
8
Comments
11 min read
What One Rust Client Can Reach Across Seven Iceberg Catalogs
xbill profile

What One Rust Client Can Reach Across Seven Iceberg Catalogs

#rust #iceberg #lakehouse #dataengineering
Comments
11 min read
Running a Jev-Style Decision Model on One TPU v6e: What Fits, What It Costs, and What Changes From a GPU
xbill profile

Running a Jev-Style Decision Model on One TPU v6e: What Fits, What It Costs, and What Changes From a GPU

#gemma #googlecloud #machinelearning #llm
10
Comments
13 min read
What Nobody Is Using in Your Azure Subscriptions, and What It Costs
xbill profile

What Nobody Is Using in Your Azure Subscriptions, and What It Costs

#azure #python #devops #opensource
2
Comments 6
19 min read
Jev After Eight Days of Independent Tests: Level With Mid-Price LLMs, Behind the Frontier
xbill profile

Jev After Eight Days of Independent Tests: Level With Mid-Price LLMs, Behind the Frontier

#ai #machinelearning #llm #gemma
4
Comments 2
17 min read
Jev After Eight Days of Independent Tests: Level With Mid-Price LLMs, Behind the Frontier
xbill profile

Jev After Eight Days of Independent Tests: Level With Mid-Price LLMs, Behind the Frontier

#ai #machinelearning #llm #gemma
11
Comments 2
17 min read
Plain Gemma 4 26B vs Jev on One EC2 L4: 2.1 Points Behind Overall, Level on Yes/No, 4.5 Behind on Multiple Choice
xbill profile

Plain Gemma 4 26B vs Jev on One EC2 L4: 2.1 Points Behind Overall, Level on Yes/No, 4.5 Behind on Multiple Choice

#gemma #aws #machinelearning #llm
1
Comments
23 min read
Plain Gemma 4 26B vs Jev on One EC2 L4: 2.1 Points Behind Overall, Level on Yes/No, 4.5 Behind on Multiple Choice
xbill profile

Plain Gemma 4 26B vs Jev on One EC2 L4: 2.1 Points Behind Overall, Level on Yes/No, 4.5 Behind on Multiple Choice

#gemma #aws #machinelearning #llm
8
Comments
23 min read
A 4 GB Laptop GPU vs a 6-Core CPU on Gemma 4, Re-Measured in ABBA Order: 4.1x
xbill profile

A 4 GB Laptop GPU vs a 6-Core CPU on Gemma 4, Re-Measured in ABBA Order: 4.1x

#gemma #llamacpp #cuda #benchmarking
10
Comments 2
11 min read
Gemma 4 on a Tesla T4, Part 2: The Minimum GCE VM and a Script to Drive It
xbill profile

Gemma 4 on a Tesla T4, Part 2: The Minimum GCE VM and a Script to Drive It

#gemma #vllm #gcp #cuda
13
Comments
13 min read
What Nobody Is Using in Your Google Cloud Projects, and What It Costs

Builds a local CLI tool to hunt ghost cloud bills
xbill profile

What Nobody Is Using in Your Google Cloud Projects, and What It Costs

#googlecloud #python #devops #opensource
12
Comments 5
14 min read
Find the AWS Resources Nobody Is Using, and What They Cost You
xbill profile

Find the AWS Resources Nobody Is Using, and What They Cost You

#aws #python #devops #opensource
Comments
9 min read
Put the Arithmetic in the Tool: an MCP Server for an AWS Waste Scanner
xbill profile

Put the Arithmetic in the Tool: an MCP Server for an AWS Waste Scanner

#aws #mcp #python #devops
5
Comments 2
8 min read
One Iceberg MCP Server, Seven Catalogs: What It Takes to Reach Each One
xbill profile

One Iceberg MCP Server, Seven Catalogs: What It Takes to Reach Each One

#mcp #iceberg #python #dataengineering
Comments 2
7 min read
One Iceberg MCP Server, Seven Catalogs: What It Takes to Reach Each One
xbill profile

One Iceberg MCP Server, Seven Catalogs: What It Takes to Reach Each One

#mcp #iceberg #python #dataengineering
11
Comments
7 min read
Gemma 4 on a Tesla T4: QAT Weights Decode 1.79x Faster Than bf16
xbill profile

Gemma 4 on a Tesla T4: QAT Weights Decode 1.79x Faster Than bf16

#gemma #vllm #cuda #machinelearning
8
Comments 1
9 min read
It's Got What Content Craves: Sanity, Built for the People of Idiocracy

Sanity Challenge Path Two Submission
xbill profile

It's Got What Content Craves: Sanity, Built for the People of Idiocracy

#devchallenge #sanitychallenge #sanity #ai
Comments 2
9 min read
Serving Gemma 4 on an AMD MI300X: What $1.99 an Hour Buys

Real-world benchmarks and MCP tool management tips
xbill profile

Serving Gemma 4 on an AMD MI300X: What $1.99 an Hour Buys

#amd #vllm #rocm #machinelearning
14
Comments 7
9 min read
An MI300X Over MCP: What the Matrix Cores Execute, and What They Don't
xbill profile

An MI300X Over MCP: What the Matrix Cores Execute, and What They Don't

#amd #mcp #rocm #machinelearning
13
Comments 5
20 min read
A 4 GB Laptop GPU Beats a 12-Core CPU by 4.3x on Gemma 4
xbill profile

A 4 GB Laptop GPU Beats a 12-Core CPU by 4.3x on Gemma 4

#machinelearning #gpu #benchmarking #python
16
Comments 7
6 min read
Four Iceberg Tools, Three Agent Frameworks: What Ports, and What Doesn't
xbill profile

Four Iceberg Tools, Three Agent Frameworks: What Ports, and What Doesn't

#iceberg #aiagents #lakehouse #dataengineering
8
Comments 1
18 min read
Four Iceberg Tools, Three Agent Frameworks: What Ports, and What Doesn't
xbill profile

Four Iceberg Tools, Three Agent Frameworks: What Ports, and What Doesn't

#iceberg #aiagents #lakehouse #dataengineering
Comments
18 min read
Nano Banana 2 Lite in Kiro CLI 3: MCP 2.0, the New Interactions API, and Headless Permissions
xbill profile

Nano Banana 2 Lite in Kiro CLI 3: MCP 2.0, the New Interactions API, and Headless Permissions

#kiro #mcp #gemini #python
3
Comments
12 min read
Nano Banana 2 Lite in Kiro CLI 3: MCP 2.0, the New Interactions API, and Headless Permissions
xbill profile

Nano Banana 2 Lite in Kiro CLI 3: MCP 2.0, the New Interactions API, and Headless Permissions

#kiro #mcp #gemini #python
9
Comments
12 min read
Build in the VM, Think on the Mac GPU: Debian 13 on Apple container With a Local Gemma 4
xbill profile

Build in the VM, Think on the Mac GPU: Debian 13 on Apple container With a Local Gemma 4

#debian #macos #ollama #linux
10
Comments 1
16 min read
Nano Banana 2 Lite, Revisited: MCP 2.0, the New Interactions API, and Three Agent CLIs
xbill profile

Nano Banana 2 Lite, Revisited: MCP 2.0, the New Interactions API, and Three Agent CLIs

#mcp #python #gemini #claudecode
11
Comments 2
12 min read
Two Rust Clients for Gemma 4: Calling the Endpoint vs. Calling the MCP Server 🦀
xbill profile

Two Rust Clients for Gemma 4: Calling the Endpoint vs. Calling the MCP Server 🦀

#rust #mcp #gemma #llamacpp
9
Comments 1
17 min read
FastMCP Is Now MCPServer on AWS: Moving a boto3 EC2 MCP Server to the MCP Python SDK 2.x
xbill profile

FastMCP Is Now MCPServer on AWS: Moving a boto3 EC2 MCP Server to the MCP Python SDK 2.x

#aws #mcp #python #ec2
Comments 1
9 min read
FastMCP Is Now MCPServer: Migrating a Python MCP Server to the MCP SDK 2.x

Dependency upgrades break shared environments
xbill profile

FastMCP Is Now MCPServer: Migrating a Python MCP Server to the MCP SDK 2.x

#mcp #python #googlecloud #vllm
14
Comments 10
12 min read
Gemma 4 on an 2021 4 GB Laptop GPU: QAT Takes It From 9.5 GiB to 1.6
xbill profile

Gemma 4 on an 2021 4 GB Laptop GPU: QAT Takes It From 9.5 GiB to 1.6

#gemma #llamacpp #mcp #cuda
11
Comments 3
13 min read
2B Gemma 4 Deployment with Cloud Run, NVIDIA L4, MCP SDK 2.x, and Claude Code
xbill profile

2B Gemma 4 Deployment with Cloud Run, NVIDIA L4, MCP SDK 2.x, and Claude Code

#mcp #gemma #googlecloud #claudecode
9
Comments 1
13 min read
Four Debian 13 Boxes, One Brief: 1,923 Packages on Metal, 328 in the Cloud
xbill profile

Four Debian 13 Boxes, One Brief: 1,923 Packages on Metal, 328 in the Cloud

#debian #linux #claudecode #devops
1
Comments 3
11 min read
Four Debian 13 Boxes, One Brief: 1,923 Packages on Metal, 328 in the Cloud
xbill profile

Four Debian 13 Boxes, One Brief: 1,923 Packages on Metal, 328 in the Cloud

#debian #linux #claudecode #devops
11
Comments 2
11 min read
The Desktop Looked Right: 2,093 Parse Errors a Boot, and 10.9 Seconds That Weren't Doing Anything
xbill profile

The Desktop Looked Right: 2,093 Parse Errors a Boot, and 10.9 Seconds That Weren't Doing Anything

#debian #linux #claudecode #devops
3
Comments 1
16 min read
The Cable Buys Headroom: 91% of a USB 2.0 Bus, 3.6% of a Thunderbolt One
xbill profile

The Cable Buys Headroom: 91% of a USB 2.0 Bus, 3.6% of a Thunderbolt One

#linux #networking #usb #debugging
12
Comments 1
18 min read
Generosity Is a Default Setting

DEV Weekend Challenge: Generosity Edition Submission 💜
xbill profile
xbill

Generosity Is a Default Setting

#weekendchallenge #linux #networking #python
2
Comments 2
9 min read
AWS Has Two Iceberg REST Catalogs: What Each One Actually Serves
xbill profile

AWS Has Two Iceberg REST Catalogs: What Each One Actually Serves

#aws #iceberg #lakehouse #dataengineering
Comments 1
8 min read
Seven Iceberg REST Catalogs: What They Declare, and What They Serve
xbill profile

Seven Iceberg REST Catalogs: What They Declare, and What They Serve

#iceberg #lakehouse #dataengineering #googlecloud
7
Comments 4
17 min read
ChromeOS Lookalikes, Two Ways: One With Drivers, One Without
xbill profile

ChromeOS Lookalikes, Two Ways: One With Drivers, One Without

#chromeos #linux #debian #gnome
8
Comments 2
8 min read
Three Gemma 4 Deployments on One T4G for Under $3: What the Runtime Changes, and What It Doesn't
xbill profile

Three Gemma 4 Deployments on One T4G for Under $3: What the Runtime Changes, and What It Doesn't

#aws #machinelearning #benchmarking #python
Comments 2
14 min read
Three Gemma 4 Deployments on One T4G for Under $3: What the Runtime Changes, and What It Doesn't
xbill profile

Three Gemma 4 Deployments on One T4G for Under $3: What the Runtime Changes, and What It Doesn't

#aws #machinelearning #benchmarking #python
8
Comments 4
14 min read
Streamline Publishing with a Claude Code Skill
xbill profile

Streamline Publishing with a Claude Code Skill

#claudecode #writing #devtools #ai
Comments 1
7 min read
loading...