About
Activity
12K followers
-
Andrew Eye reposted thisAndrew Eye reposted thisWhere are y’all storing all the markdown documents your teams generate every day to give coding agents context? Git? A collection of other tools? And how are you connecting those documents to the work that actually gets shipped? We’re working on this at Closedloop.ai: a durable system of record for that documentation, with version history, collaborative editing, and comments for review. Teams can work together in an editor while the platform handles conversion to and from markdown. Agents can read and write documents and comments through MCP. Documents also have workflows and can be linked together, connecting the context behind a feature with its work status. Now connect those documents to your team’s agent sessions and the resulting PRs. You can trace a feature from the documents that define it through the sessions that implement it to the code changes produced. You can also aggregate token usage and spend across that work, alongside code churn, to understand the AI usage cost of delivering a feature. We already have full session and PR tracking available today. We’ve also been using the platform to manage our own documentation internally for months, and we’ll be releasing those documentation capabilities publicly soon. We’re looking for early adopters who want to try this and give us candid feedback about what works and what doesn’t. If this sounds useful to your team, reach out and we can set up a demo. It won’t be a sales call—we don’t currently have anything to sell you. We want to learn from how other teams work.
-
Andrew Eye shared thisPumped for our next AI Meet up! Daniel Ochoa will be presenting closedloop-intel, a tool our team built to connect tickets, plans, branches, and code into one queryable graph. It gives engineers and AI agents the context to answer questions like “is someone already building this?” and “what actually shipped last week?”. Come join us Tuesday at the Capital Factory at 6:30pm! RSVP - https://luma.com/iwe60d9cAndrew Eye shared this"If I change this file, whose work breaks?" Every engineer has asked some version of that and gotten no clean answer, because it's scattered across your tickets, your repo, your branches, and three Slack threads. No single tool sees across all of it. At our next meetup, Daniel Ochoa is showing closedloop-intel, something our team built to answer exactly those questions in milliseconds: what breaks if you change this, whether someone's already building it, what actually shipped last week. It also hands your AI agents the org context they usually waste tokens piecing together. Also, one more speaker still to come! RSVP linked below. 📆 Tuesday, September 22 at 6:30pm 📍STATION Austin 👉 RSVP here - https://luma.com/iwe60d9c
-
Andrew Eye shared thisOne of the more interesting things happening as AI takes on more of the actual building is that everyone's job is starting to bleed into everyone else's. That's a big part of what we're talking about tomorrow night at STATION Austin. Kaiti Carpenter, our Head of Product at ClosedLoop, is going to talk about what the PM role looks like when AI suddenly gives one person the ability to do work that used to be spread across design, research, data, QA, and GTM. Her talk is called "Concentrate: The Too-Sweet New PM Role (& Ways to Bruce Lee It)." If you know Kaiti, you know the title makes sense. Matthew Stephens is coming at a similar shift from the design side. He's built a Behavioral Agentic Design System (BADS) around the idea that if agents are going to build interfaces, your design system has to teach them more than which components to use. It needs to give them context about how and why your team designs the way it does. I think these are two sides of a pretty important conversation about what happens to product teams when AI can actually do more of the work. Come argue about it with us tomorrow. Tuesday, August 25 · 6:30 PM STATION Austin / Capital Factory RSVP: https://lnkd.in/gdYRSXfA
-
Andrew Eye shared thisIf you're experimenting with AI in your design workflow but aren't happy with the results, come to our meetup tonight. We'll demo the prototype sandbox we use internally. An agent takes a plain-language idea and turns it into a full, clickable page using our real design system, not a throwaway Figma mockup.If you've been fighting AI prototypes that drift away from your design system or need constant cleanup, it's worth seeing. We'll also cover the Claude Code fundamentals our engineers use, including how we measure whether an agent session is delivering value or just burning tokens. We're hiring designers, too. If you're building with AI and want to see how we work, come find me tonight. The RSVP link and our open design role are both in the comments. 📅 Tuesday, July 28 · 6:30 PM 📍 Capital Factory (STATION Austin), 701 Brazos St, Austin #AI #Design #ClaudeCode #Austin #Hiring
-
Andrew Eye shared this"Lines of code" was always a weak way to measure engineers. Now that agents write most of the lines, it's close to useless. The mistake is thinking that means you stop measuring. You don't. You change what you measure. Our team started looking at something closer to value over cost per session: what actually got merged, against what it cost in tokens and human attention. Some sessions look incredibly busy and ship nothing. Some quiet ones are wildly efficient. You can't tell which is which by watching the diff scroll by. That gap between looking productive and being productive is where I think most teams are flying blind right now. We're getting into it at our next meetup Tuesday 7/28. Mike Angstadt will walk through what a good agentic coding session actually looks like, then we'll show how we measure it. 📅 Tuesday, July 28 · 6:30 PM 📍 Capital Factory (STATION Austin), 701 Brazos St, Austin 🎟️ RSVP: https://luma.com/563dzwzo #AI #ClaudeCode #SoftwareEngineering #Austin
-
Andrew Eye shared thisOur VP eng. and highest contributing engineer are delivering a free online course on mastering Claude Code Monday at 9AM CST. There's 5 hours of deep content on context management, token optimization, skills development and more. The agenda is published so sign up and drop in for whatever portions seem relevant. Only 100 spots available so sign up quick!Andrew Eye shared thisWe sped up software delivery roughly 16x in the last year. On Monday, we're showing exactly how. We're opening up a free, hands-on Claude Code training session. Anyone can join. It's led by Mike Angstadt, our VP of Engineering, and built around the operating patterns our engineers use in production every day, not a documentation readout. Across five modules you'll cover building Claude Code primitives, planning and context management, intent recovery with evidence, review and verification, and designing a repeatable workflow. You'll leave with seven small artifacts you can reuse at work. You don't need to attend the whole day. Each module is self-contained, so join for the topics that matter most to you. Monday, June 22 · 9:00 AM to 3:30 PM CT · live on Google Meet Save a spot: https://lnkd.in/grTvnbvD
-
Andrew Eye shared thisIf you’re leading or building on an AI-first software team, tonight’s meetup will probably feel very familiar. Most teams are seeing the same thing happen: AI dramatically increases output at the individual level. But at the team level, new problems start showing up... - review bottlenecks - fragmented context - hidden decisions - unclear ownership - rising token costs - more work happening in parallel than teams know how to coordinate So tonight we’re discussing two topics that are becoming hard to ignore: 📊 Beyond DORA - The New Metrics for AC/DC (Agentic Coding Delivery Cycle) How should teams measure engineering performance when agents are integrated into the workflow? 💸 Are You Ready for Metered Token Billing on June 15? What considerations should startups have regarding the economics of headless agents and usage based execution? This isn't a "future of AI" panel. It's an opportunity for engineers, founders, PMs, and technical leaders to share insights on what is effectively working today. RSVP Here - https://luma.com/n5b0ds7m
-
Andrew Eye shared thisBIG NEWS! My wife Kate has spun out her Risk and & Advisory practice and is running her own business again! Know someone who needs a SOC2 or other Risk & Compliance work? Please reach out! Tell them you want the #AEYEDISCOUNT. ;-) Go get em Kate! So proud of you and excited to watch you grow this business even further! Read on below... Kate's version is way better!Andrew Eye shared this🥁Drum roll....I have a BIG announcement 🎉🥳 (and a small ask). I'm thrilled to announce the launch of Eye & Co - a modern, tech-forward assurance and advisory firm offering SOC 1 / SOC 2 examinations, regulatory compliance, tech diligence, and AI risk & governance services. Come see what we've built at www.eyecoadvisory.com! 🙌 The Eye & Co Dream Team includes Ryan Kline Sabol, Zada Villalobos, and Zachary Sly. Huge thanks to each of them for joining me on this adventure. 🙏 And now for the ask...we are starting with an amazing team and a solid client base, but we have a lot of growing and learning to do. I'm generally not one to ask for help, but my husband and favorite person Andrew Eye is a big fan of the "humble startup card," so I'm going to give it a shot. Here goes: ⭐ To my friends, clients, and colleagues - we are but a humble startup and we would LOVE any referrals, insights, words of wisdom, website critiques, shares, follows, mentors, client testimonials, and whatever else we need and didn't even think of. ⭐ Lets chat or grab a coffee on me - DM me or email at kate@eyecoadvisory.com. Cheers!Eye & Co | Tech-Forward Advisors for Compliance & SecurityEye & Co | Tech-Forward Advisors for Compliance & Security
-
Andrew Eye shared thisWe’re hosting the second Full Lifecycle AI Software Teams Meetup next Tuesday (5/26) at 7pm and I’d love to see you there. The first one was awesome. We had a great group of founders, engineers, PMs, designers, and technical leaders come out to talk about how AI is changing the way software actually gets built. I shared a bit about our own journey at ClosedLoop.ai and how our team has gotten dramatically faster with AI. The biggest lesson for us has been that the hard part is not just writing code faster. It’s figuring out how the whole team plans, coordinates, reviews, tests, and ships when the pace of work changes. The best part of the night was the discussion after. Lots of great questions, lots of people comparing notes, and a bunch of really interesting conversations about what is working, what is still messy, and where this is all headed. That’s exactly the kind of energy we want to keep going. For the next meet up, we’ll have two AI-related talks followed by Q&A and networking. If you’re building software, leading a product or engineering team, experimenting with AI in your workflow, or just curious where this is going, come hang out with us. RSVP here: https://lnkd.in/gDNX_aYg Hope to see y’all there!
-
Andrew Eye liked thisAndrew Eye liked thisWhere are y’all storing all the markdown documents your teams generate every day to give coding agents context? Git? A collection of other tools? And how are you connecting those documents to the work that actually gets shipped? We’re working on this at Closedloop.ai: a durable system of record for that documentation, with version history, collaborative editing, and comments for review. Teams can work together in an editor while the platform handles conversion to and from markdown. Agents can read and write documents and comments through MCP. Documents also have workflows and can be linked together, connecting the context behind a feature with its work status. Now connect those documents to your team’s agent sessions and the resulting PRs. You can trace a feature from the documents that define it through the sessions that implement it to the code changes produced. You can also aggregate token usage and spend across that work, alongside code churn, to understand the AI usage cost of delivering a feature. We already have full session and PR tracking available today. We’ve also been using the platform to manage our own documentation internally for months, and we’ll be releasing those documentation capabilities publicly soon. We’re looking for early adopters who want to try this and give us candid feedback about what works and what doesn’t. If this sounds useful to your team, reach out and we can set up a demo. It won’t be a sales call—we don’t currently have anything to sell you. We want to learn from how other teams work.
-
Andrew Eye liked thisAndrew Eye liked this"If I change this file, whose work breaks?" Every engineer has asked some version of that and gotten no clean answer, because it's scattered across your tickets, your repo, your branches, and three Slack threads. No single tool sees across all of it. At our next meetup, Daniel Ochoa is showing closedloop-intel, something our team built to answer exactly those questions in milliseconds: what breaks if you change this, whether someone's already building it, what actually shipped last week. It also hands your AI agents the org context they usually waste tokens piecing together. Also, one more speaker still to come! RSVP linked below. 📆 Tuesday, September 22 at 6:30pm 📍STATION Austin 👉 RSVP here - https://luma.com/iwe60d9c
-
Andrew Eye liked thisAndrew Eye liked thisOn Tuesday, August 25, we're back at Capital Factory (STATION Austin) with two great talks from people shaping how AI-native products are built. Matthew Stephens (fractional design & AI leader, ex-OneSpot, DeviantArt) will introduce BADS (Behavioral Agentic Design System). As agents take on more of the product development process, design systems need more than tokens and components. They need opinions, ethics, accessibility, and memory so agents produce coherent, on-brand experiences instead of the same generic dashboards. Matthew built BADS with Claude Code, and he'll show how it works. We'll also hear from Closedloop's Head of Product, Kaiti Carpenter (more details coming soon). Same format as always: two talks, Q&A, food, and plenty of time to connect with other people building AI-native teams. 📅 Tuesday, August 25 · 6:30 PM 📍 Capital Factory (STATION Austin), 701 Brazos St, Austin 🎟️ RSVP: https://lnkd.in/gaYbZtju 🚗 Parking: Validation is available at the Omni Hotel garage between 5:00 PM and 9:30 PM. Enter after 5:00 PM and pick up a validation coupon from the STATION Austin front desk to reduce parking to $8. Street parking is also free after 6:00 PM on Tuesdays. #AI #ClaudeCode #SoftwareEngineering #DesignSystems #Austin
Experience & Education
Volunteer Experience
-
Board Member
Autoimmune Hepatitis Association
- Present 6 years 2 months
Health
Autoimmune hepatitis (AIH) is a rare chronic disease that occurs when patients’ immune systems attack their livers and sometimes damage them. With treatment, most patients don’t need liver transplants and have normal life expectancies.
In July 2017 my daughter was diagnosed with AIH. Thanks to her wonderful doctor she's been doing great for 3+ years. In 2020 I was fortunate enough to join the board of the AIHA. If you are interested in helping the cause, please reach out. -
Board Member
Service Dogs, Inc.
- 2 years
Health
Service Dogs, Inc. transforms dogs abandoned to animal shelters into lifelines for people living with hearing loss or mobility challenges. Founded in 1988 by attorney Sheri Soltes, SDI provides the dogs and training free of charge.
-
Cat Herder / Coach
Lonestar Soccer Club
- Present 26 years
Children
I'm a lifelong soccer player and have been fortunate enough to coach youth soccer for nearly 20 years. The highlight of this experience has been coaching my own daughters and their friends.
Patents
-
Hero cards that display contextual information and actions for backend systems
Issued US 10560416
Systems herein supply hero cards that display contextual information and actions based on backend systems. The information and actions are related to a trigger content item in an application that does not have access to the backend systems. A hero agent running a user device can communicate the trigger content to a hero server. The hero server can determine which connectors to backend systems relate to the user device, and then can retrieve result information from one or more of the backend…
Systems herein supply hero cards that display contextual information and actions based on backend systems. The information and actions are related to a trigger content item in an application that does not have access to the backend systems. A hero agent running a user device can communicate the trigger content to a hero server. The hero server can determine which connectors to backend systems relate to the user device, and then can retrieve result information from one or more of the backend systems. The result information populates the hero card, which the hero agent displays within the application. Action buttons on the hero card cause the hero agent to contact one or more of the backend systems to cause an action to take place there.
Other inventorsSee patent -
GESTURE-BASED WORKFLOW PROGRESSION
Issued US US20140282243A1
See patentMethods, systems, and apparatus, including computer programs encoded on a computer storage medium, for displaying, by the one or more processors, at least one graphical representation associated with an electronic document on a touchscreen display; and receiving, by the one or more processors, user input to the at least one graphical representation, the user input indicating a touchscreen gesture to the touchscreen display, and in response: determining a type of the touchscreen gesture;…
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for displaying, by the one or more processors, at least one graphical representation associated with an electronic document on a touchscreen display; and receiving, by the one or more processors, user input to the at least one graphical representation, the user input indicating a touchscreen gesture to the touchscreen display, and in response: determining a type of the touchscreen gesture; determining a degree of the touchscreen gesture; identifying one or more actions based on the type and the degree, each action of the one or more actions being executable to progress through a workflow; and displaying one or more action elements on the touchscreen display, each action element representing an action of the one or more actions.
Recommendations received
10 people have recommended Andrew
Join now to viewOther similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top contentOthers named Andrew Eye in United States
5 others named Andrew Eye in United States are on LinkedIn
See others named Andrew Eye