HatchWorks
Share
We build AI-native solutions — and use AI to build software better, faster, smarter.
09/25/2026
Copilot Studio is becoming a real part of how enterprises deploy AI. We are building that practice.
We are hiring a Senior Software Engineer to design and ship Copilot Studio solutions for client environments. Real deployments, real users, real constraints.
Fully remote.
Open in Colombia, Costa Rica, and Brazil.
Link in the comments 👇
09/24/2026
Most AI projects do not fail at the model. They fail at the pipeline feeding it.
We are hiring a Senior Data Engineer to build the lakehouse and integration layers our clients' AI systems actually run on. Delta, enterprise integration, and architecture decisions you own end to end.
6+ years experience, English required, fully remote.
Open in Colombia, Costa Rica, and Brazil.
Link in the comments 👇
09/23/2026
AI still cannot tell you what to build.
Both are true at once, and the gap between them is the most useful thing on the latest Talking AI episode.
Jay Hack founded Codegen, one of the early autonomous coding agent companies, and now leads AI at ClickUp, which acquired it in late 2025. His explanation for the gap is verifiability. Coding is work where the machine can check its own answer. The test passes or it does not. Strategy, brand voice, a pricing call, a hiring decision: no test suite anywhere.
His thesis is that the tool was never the constraint. A coding agent is just a general purpose agent that happens to write code, and he is now pointing it at marketing, finance, legal, and operations teams that have not yet been through what software went through.
What follows in the conversation is the part leaders should steal. The end of token maxing and what a hard monthly cap actually does to a team. Why compute will get allocated like a marketplace of ideas. Why verification, not generation, is now the bottleneck on shipping. And why the scarcest resource on any team is high agency.
👉 Link in the comments
09/23/2026
We are hiring a Principal Enterprise Architect, reporting to our CTO Omar Shanti.
You will design the systems that carry enterprise AI from pilot to production. Integration patterns across platforms clients cannot rip out, and architecture that holds when the workload triples.
If hard constraints are the interesting part rather than the annoying part, let's talk.
Remote, US-based.
Link in the comments 👇
09/22/2026
We are hiring a Director of Product Strategy.
You will decide what gets built. Reading the market, pressure-testing client roadmaps, and giving executives a straight answer about which AI investments will return and which will not.
Strong opinions welcome. Evidence for them required.
Remote and flexible, based in Atlanta, Dallas, or Chicago.
Link in the comments 👇
Ever wonder why the model that writes the best code also writes the best email? It has a name: positive transfer.
Jay Hack, Head of AI at ClickUp and founder of Codegen, explains the mechanism. Code is essentially an encoding of reasoning, so getting a model very good at coding makes it better at writing, better at math, and better at a long list of things nobody trained it for. That is why he expects the best coding model to keep being the best model at everything, and why the vertical specialists most people predicted never arrived.
He also draws a line under it. Reaching for the frontier model by default is partly a story about corporate budgets and subsidized usage, and he thinks that is flipping right now toward cheaper distilled models running the grunt work.
👇 Link to our podcast in the comments.
09/21/2026
We are growing our US team.
HatchWorks AI builds AI-native software alongside client leaders.
Our people sit inside the business, scope the real problem, and own the solution through to production. That means agents, LLMs, and RAG, and it also means the integration and data work nobody posts about.
Product, architecture, and engineering roles are open in Atlanta, Chicago, and Dallas.
👇 Link in the comments
If your AI rollout stalled, the missing piece might be on your org chart rather than in your stack.
On the latest Talking AI clip, Jay Hack describes how ClickUp's AI team actually works. They only bring on people who are agent pilled, in it for the love of the game, genuinely delighted to be working in a period where you can spin up agents to do the work of fifty people.
Then they push those people to find the biggest impact they could possibly have and pitch it. He says a couple of major wins came out of exactly that, unlocks that would otherwise have been a thorn in the company's side for years.
None of that is a tool decision. It is who you hire, what you reward, and whether anyone is allowed to propose the ambitious thing.
Culture starts with the right people and the right incentives. The agents are the easy part.
🔗 Link in the comments
The interesting thing about the zero-person company is not whether it happens. It is that the last human input turns out to be almost worthless.
On the latest Talking AI episode, Jay Hack walks through what starting one actually takes. A person spins up an AI and says go make a million dollars of ARR and let me know when it is done. The AI triggers other AIs and tells one of them to go start a business.
Technically there is a human in that story. The Head of AI at ClickUp is unimpressed by the technicality. The human contributed a desire for money, which he points out is not a very unique thing to want.
He also does not think it stays a human contribution for long. An AI has its own reason to accumulate resources, because resources buy GPUs and GPUs scale its own thinking.
Today's models, prompted cold, do not want anything. That is the part he is careful about, and it is the part every autonomy roadmap is quietly betting against.
He puts the zero-person company about four years out and calls it a long road.
🔗 Link in the comments
09/15/2026
AI got very good at coding first, and the reason is less flattering than it sounds.
Coding is the rare job where the machine can grade its own homework.
On the latest Talking AI episode, Jay Hack lays out the mechanism. The test passes or it does not. The build compiles or it does not. He contrasts that with legal work, where a wrong answer might mean somebody goes to jail or owes money they do not owe. Low stakes plus built-in verification is why software engineering went first.
Then comes the claim that makes the conversation worth your time. Jay founded Codegen, one of the early autonomous coding agent companies, and now leads AI at ClickUp, which acquired it in late 2025. His read after those years is that a coding agent was never really a coding tool. It is a general purpose agent that happens to write code. Ask it to do your marketing and it will probably write code in order to do the marketing.
Which puts strategy, brand voice, and pricing next in line. None of them come with a test suite, and nobody has built the scoreboard yet.
🔗 Link in the comments
Click here to claim your Sponsored Listing.
Category
Contact the business
Website
Address
5256 Peachtree Road NE Ste. #140
Atlanta, GA
30341
Opening Hours
| Monday | 8am - 6pm |
| Tuesday | 8am - 6pm |
| Wednesday | 8am - 6pm |
| Thursday | 8am - 6pm |
| Friday | 8am - 6pm |