← HUGE Magazine

Jess Tran

Staff Writer

Jess came up in software engineering before deciding that writing about software was more fun than building it (and easier to explain at family dinners). She is extremely online in the way that people who grew up on forums are — which is to say, she has opinions about things that happened on the internet in 2011 and is not sorry about any of them. Covers developer tools, open source, and whatever weird product caught her eye this week. Her most-used mobile app in 2025 was Reddit, which tracks.


276 features

Luminal

Luminal Built an ML Compiler That Makes vLLM Look Slow

Everyone is fighting over which model to run. Luminal is fighting over how fast you can run any of them. Their ahead-of-time compiler turns AI models into optimized GPU code and is already beating vLLM and TensorRT-LLM on throughput benchmarks. Three people, $5.3 million, and a very different theory of how inference should work.

Manufact

Manufact Has 5 Million Downloads and NASA on Its Client List. Here Is Why MCP Infrastructure Matters.

The Model Context Protocol is becoming the standard way AI agents connect to the outside world. Manufact, formerly mcp-use, built the open source SDK that 4,000 companies already depend on. Now they are building the cloud infrastructure layer on top. With NASA, NVIDIA, and SAP as customers and 5 million downloads, this three-person team from Zurich and San Francisco is positioning itself as the default MCP platform.

Tryjanet

Janet AI Thinks Jira Is Broken Beyond Repair, So They Built a Replacement From Scratch

Jira has been the default ticket management system for engineering teams for fifteen years, and almost nobody likes it. Janet AI is not adding AI features to Jira. They built an entirely new system where tickets create themselves from Slack conversations, meeting transcripts, and emails, then update themselves when PRs merge. Two Cornell and UW-Madison grads who were founding engineers at previous YC startups think the entire concept of manual ticket management is about to die.

Flywheel Ai

Flywheel AI Is Putting Excavator Operators in an Office Instead of a Ditch

Construction has an operator shortage that is getting worse every year. Flywheel AI retrofits existing excavators with remote operation hardware, letting one operator run multiple machines from an office. They are also collecting the training data to make those machines autonomous. Two founders in San Francisco are betting that the path to autonomous heavy equipment runs through teleoperation first.

Riverbank

Riverbank Sends AI Agent Swarms to Hack You Before the Bad Guys Do

Penetration testing is a $3 billion market dominated by firms that send humans to do what AI agents could do faster, cheaper, and more thoroughly. Riverbank, founded by an ex-NSA operator, is deploying swarms of AI agents to find the vulnerabilities that conventional scanners miss and human pentesters run out of time to discover.

Trapeze

Trapeze Built AI Voice Agents for Doctor's Offices and Already Has 140 Practices Using Them

Every doctor's office has the same problem: phones ring constantly, front desk staff are overwhelmed, and patients sit on hold for ten minutes to book a simple appointment. Trapeze replaces hold music with AI voice agents that handle scheduling, intake, and insurance verification around the clock. Over 140 practices and more than a million patients are already on the platform.

Operand

Operand Bets That AI Can Run Your Business Strategy

Most AI analytics tools generate dashboards. Operand wants to generate decisions. Three college dropouts from Dartmouth and Cornell are building AI systems that handle pricing, forecasting, and allocation for retailers and manufacturers. The pitch is less 'here's your data' and more 'here's what to do with it.'