Why this role exists
neatlogs helps teams understand why their AI agents fail in production, then fix what went wrong. We see the messy, specific cases behind that work: an agent that follows its instructions and still gets a customer’s problem wrong; a trace that looks fine until someone from support reads it; a fix that helps one failure and creates another.
There’s a lot to say about this, but the useful parts rarely arrive as ready-made posts. They’re in customer calls, product reviews, traces, experiments, and conversations with the team.
We have a voice, an educational video series, product updates, and an active presence on X and LinkedIn. We need someone to find the strongest ideas in that work and develop them at the depth they deserve: sometimes a post, sometimes a detailed guide or playbook that a team can use. Then we need to get that work in front of the right people and learn what it helped them do.
What you’ll work on
- Blogs and long-form content. Own a small, high-quality pipeline of technical articles, guides, case studies, and product stories. Find the question worth answering, interview the people who know, research what’s missing, write the piece, and work with engineers to make it accurate. Use real runs and examples when an abstract claim won’t do.
- Playbooks and gated resources. Build practical assets a team can use to improve an agent: for example, a guide to investigating a failed run or a playbook for setting up evals. Decide when a resource should be open and when it’s valuable enough to ask for an email. Make the asset useful on its own, then give people a clear next step if neatlogs can help.
- Ajay’s voice and neatlogs’ voice. Work closely with Ajay, our founder, on ideas for his accounts, and own the company’s social channels. They shouldn’t sound identical. Learn what he would actually say, push back on weak angles, and write posts he can stand behind.
- Distribution and follow-through. Turn the best long-form work into posts, threads, short video scripts, and useful replies. Help people find it through search, relevant communities, and the channels we already use. Stay in the conversation after publishing and use the questions people ask to improve the piece or make the next one.
- A sharper sense of what works. Pay attention to who reads, shares, responds, signs up, or asks to talk. Work out which pieces bring in teams we can help and what those teams needed to understand first. Use that evidence to change a format, update an older guide, or stop a series. Reach matters, but it isn’t the only result.
Your first few months
- Month one: Get close to the product and the people using it. Review what we’ve published, join customer conversations, start shipping posts on the founder and company accounts, and choose the first substantial piece worth making.
- Months two and three: Publish a researched guide or playbook and build a distribution plan around it. Keep a manageable social rhythm, help produce the next educational pieces, and show us what you learned from the response.
- After that: Own an editorial pipeline of articles and resources that answer real questions from agent teams. Keep the strongest pieces current and use what readers and customers tell us to decide what comes next.
What good work looks like
An engineer shows you a run that failed. You ask enough questions to understand what went wrong, why it mattered to the user, and what changed. You turn it into a guide with the evidence and decisions intact, then find the shorter stories inside it for X, LinkedIn, or video. Each format does its own job. The guide helps someone solve a problem even if they never sign up.
You don’t need a detailed brief to find an angle. You can also tell when an idea isn’t ready, ask for the missing evidence, or say that a post sounds more impressive than the underlying work warrants.
After publishing, you stay in the conversation. You notice the questions people ask, see whether the right teams found it, and use that to improve the resource or make the next piece better.
What we’re looking for
- You can write clearly and edit hard. You can make a technical point understandable without flattening what makes it interesting.
- You can research, structure, and finish a substantial piece of writing. You know how to use interviews, examples, screenshots, and source material to make it more useful than a summary of what others have said.
- You’ve made content for people who build or use software. Show us both a longer piece and social work, what was yours, and what happened after each went out.
- You have good instincts for X and LinkedIn: what deserves a short post, what needs a thread or video, and when a reply is more useful than another announcement.
- You understand the basics of search intent, distribution, and conversion. You can judge whether a guide is reaching and helping the people it was made for.
- You can interview an engineer or customer, find the detail that matters, and turn it into a piece of work without putting words in their mouth.
- You’re comfortable working with a founder who has opinions, giving your own, and changing direction when the facts change.
| We look at | We don’t weigh heavily |
|---|---|
| The work you personally wrote, edited, or produced | Follower counts without context |
| The judgment behind your angles and edits | Publishing volume for its own sake |
| Whether the work helped the intended reader | Viral posts unrelated to the audience we serve |
Experience with developer tools or AI agents helps. Curiosity, accuracy, and the ability to learn fast matter more.
The working reality
This is a hands-on role. You’ll write a lot, ask a lot of questions, and switch between planned long-form work and something that just happened in the product or the community. Some days a good post will come from an hour with an engineer; other days a guide will take weeks of research and editing. A promising idea may fall apart once you check the trace.
You won’t be asked to produce every format every week. Part of the job is choosing which ideas merit that time and keeping the rest of the work moving at a sustainable pace.
You’ll work closely with Ajay, our technical team, and the people making videos and educational material. You’ll own the content and social program, with the Marketing Head setting broader marketing priorities.
This probably isn’t the right role if you mainly want to manage a content calendar, outsource the writing, or publish search-driven articles without understanding the product or the reader.
About neatlogs
neatlogs helps teams understand why their AI agents fail in production, fix the cause, and check whether the fix worked. It connects production traces with the judgment of the people who know what the agent should have done, then helps the team investigate what happened and decide what to change.
Apply for this role
Tell us about yourself and share work you’re proud of. A real person reads every application.