# Hackathon Judging Feedback: 5 Teams I Judged at LA Hacks AI > My hackathon judging feedback from Round 1 of the LA Hacks AI Hackathon 2026: five teams, four with no slides, and what each one should have done instead. Canonical: https://thehackathonplaybook.dev/blog/hackathon-judging-feedback-la-hacks-ai Last updated: 2026-11-05 --- ## Most of These Teams Lost Me on Communication, Not Code Code wasn't the problem for any of these teams. I judged Round 1 at the first LA Hacks AI Hackathon, at UCLA's James West Alumni Center on October 18, 2026. Of the first five teams I took notes on, **four had no slides or visuals of any kind**. Two left me unsure what the project even did. And one team with a very good product and demo still had a problem, which I'll get to. Most of it is easy to fix: describe your product in **1 or 2 sentences**, put **a slide** up, and be ready before the judge walks up. | Hackers | Judges | Per Table | Had No Slides | | --- | --- | --- | --- | | ~500 | ~110 | 5-7 min | 4 of 5 | > **How Round 1 Worked:** Round 1 was science-fair judging. Each judge visited about 15 tables for 5 to 7 minutes each, then ranked every project they saw. The top 5 went on to Round 2: a 3-minute pitch, 4 minutes of Q&A, and a ranked-choice vote by all the judges. The rubric scored Technical Depth, Product Thinking & Impact, Originality & Insight, and Execution & Polish. Here are the five teams, in the order I judged them. No names, just what happened and what I'd tell them. ## Team 1: I Gave It a 0 I walked up to the table, and the hacker asked me to wait until they finished eating. When they did start, there was no page, no slide, nothing to set up the project. They went straight into the demo, then showed me a video after it. I was so lost. I didn't know what the project was doing or what problem it was tackling. I gave it a 0. > **The Rubric Doesn't Go to 0:** The lowest score on the LA Hacks AI Hackathon 2026 judging rubric is a 1. For Product Thinking & Impact, a 1 means "Unclear: vague problem. Weak user definition." If I walk away not knowing what your project does, that's where you end up. **Be ready when the judge walks up.** Have the laptop open, the demo loaded, and a slide up that says what the project is. ## Team 2: I Missed Half the Pitch Overall, this one was pretty good, but there were a couple of issues. First, the pitch was spoken only. The demo room was very loud, and I couldn't follow about half of it. A slide deck I could follow along with while they talked would have made it **100 to 200% better**. The demo was OK. Then I asked about the project's future, and they said it was "out of scope." They deflected. > Never say out of scope. Never say we'd never consider. That's just a very bad look. It shows you don't have future views on your project. > > Bill Zhang, Round 1 judge at the LA Hacks AI Hackathon 2026, from my judging notes You should definitely have some considerations for that. Even if you haven't built it, have an answer for what you'd do next. ## Team 3: Great Engineering, No Answers Ready Technically, this one was really interesting, and a solid project. Then I asked follow-up questions, and they had a hard time answering. No visuals, no appendix slides. They answered everything off the cuff and looked things up by hand while I waited. They also weren't thinking long-term. I asked how they'd take it to production, and they said it was "just a proof of concept." Same thing as Team 2. **One good slide on how they'd productionize it would have helped a lot.** **Appendix Slides Worth Having for Table Judging** - [ ] How you'd take it to production - [ ] What comes next, and who would use it - [ ] An architecture diagram for technical judges - [ ] Answers to the questions you know are coming ![Architecture slide showing a 911 call audio data pipeline, transcription, cleanup, transformation, fine-tuned model, frontend dashboard, backend server, Twilio, Retell, and Hume emotional analysis.](https://thehackathonplaybook.dev/blog/berkeley-ai-hackathon-appendix-architecture.png) *An appendix slide from our Dispatch AI pitch at the UC Berkeley AI Hackathon 2024. When a judge asks how it works, you pull this up instead of searching for the answer.* So, solid project. It just wasn't polished enough, and they couldn't answer my follow-up questions in a coherent and organized way. Appendix slides would've helped a lot. ## Team 4: A Very Good Product With a Delivery Problem This is the team I mentioned up top. The pitch was very nice. The product was very good, and so was the demo. The problem was soft skills. The presenter wasn't really looking me in the eye, and kept rambling for a long time. I got kind of confused, and I couldn't really ask my questions. That's about it. The product and the demo were very good. It's just how they pitched that was the issue. > **Leave Room for the Judge's Questions:** If I can't ask my questions, I can't figure out what I missed. Look at the judge, keep your answers short, and leave room for questions. ## Team 5: He Talked So Much I Got Lost This guy talked so much. There were no slides or visuals of any kind, and I was getting so lost. I couldn't tell what the project was supposed to do. Your pitch should be **3 to 4 sentences**. You should be able to describe your product in **1 or 2 sentences, max**. If you can't describe your project in 1 or 2 sentences and make it obvious what it does, you need to find a way to do that ASAP. Or at least use some sort of pitch deck. [Pitch Script and Judge Q&A Prompts](https://thehackathonplaybook.dev/cheat-sheet#pitch): Paste-ready prompts that help you write the two-sentence version of your project and rehearse the follow-up questions judges ask. ## What I'd Tell Every Team Before Round 1 Most of what I saw came down to how teams explained their project. Here's what I'd tell every team: **Do:** - Be ready before the judge walks up - Describe the product in 1 or 2 sentences - Keep the pitch to 3 or 4 sentences - Put a slide up, especially in a loud room - Have appendix slides for follow-up questions - Make eye contact and leave room for questions **Don't:** - Asking the judge to wait while you finish eating - Jumping straight into the demo with no setup - "It's out of scope" - "We'd never consider that" - "It's just a proof of concept" - Talking until the judge is lost Before judging starts, figure out how to describe your project in 1 or 2 sentences, and put up a slide that says what it is. Have a slide ready for what's next, too. [The Pitching Playbook: Table Judging](https://thehackathonplaybook.dev/playbook/pitching): The full pitching phase, including the table-judging section built from this round: being ready, the two-sentence description, the one slide to bring, appendix answers, and what to say instead of "out of scope." [9 Hackathon Pitch Mistakes I Saw at LA Hacks](https://thehackathonplaybook.dev/blog/hackathon-pitch-mistakes-la-hacks): From judging the main LA Hacks event in Pauley Pavilion: pitching every judge the same way, skipping the backup video, and leaving no time for Q&A.