Frontier engineering with a philosophy attached.
For people who want the 'how' to matter as much as the 'what' — and who'd rather work somewhere with products they can point to, not just clients they can't name.
The same four principles we build systems on, apply to how we work.
Context before capability
A system models who it serves, who it affects, and what happens when it is wrong — before it is permitted to act. Not a safety review after it's built.
Wellbeing is the metric
Success is whether the human is better off. Not engagement. Not retention. Not task volume.
Judgment stays human
Every system is built to make a person better at deciding, not to quietly absorb the decision.
Explainable under pressure
Explainable to the person who relies on it, the board that governs it, and the regulator who eventually asks.
Every project runs on Conscipact HAI — Human + AI Intelligence. You’ll use AI to carry volume, drafts, and coverage, but the judgment calls stay yours, and your name is on what ships. It’s a different rhythm than a shop that either bans AI tools or lets them make the decisions — here, the division of labor is deliberate, and it’s the same discipline you’ll see in the products we build. People who already work here have written about it, in their own words, at Stories from Real Humans.
3 openings, right now.
Every posting here is live from our own hiring system — if it’s on this page, it’s open.
Human-AI Software Developer (Mid)
Job ID COJ00219Build and run production software where an AI model sits inside the flow of somebody's work or life, and own the parts that decide what the model is allowed to do, what a human has to confirm, and what happens when the model is wrong.
Read the role and apply →Human-AI QA Engineer (Mid level)
Job ID COJ00220You will test AI systems where the correct output is not a fixed string: writing evaluation suites for model behaviour, building regression harnesses that catch drift between model versions, and finding the failure cases that only show up under real user pressure. You sign off on what ships.
Read the role and apply →Human-AI Software Engineer - Intern
Job ID COJ00221Write, test and ship parts of Conscipact's own AI products alongside senior engineers, working with AI agents that carry the volume while you carry the judgment on what actually goes out. Two-month unpaid internship in Hyderabad, hybrid, with full-time employment possible after it depending on your work and on openings at that time.
Read the role and apply →Nothing here that fits?
Then walk through the Open Door. We don’t always have a posting for the work you do — that’s a fact about us, not about you.C.O is Conscipact’s own AI agent, built by our team. Leave your details, and it may suggest your profile when a relevant role opens. The experience you tell us counts forward, so a role that opens next year considers the experience you’ll have then.
Open the doorAlready applied, or already left your details here? Check where it stands with the reference we emailed you. Anything else, write to reach@conscipact.com.
What happens after you apply
You get a reference number when you apply. C.O, Conscipact’s own AI agent, may help assess your application against the role, and you can use your reference to check the latest recorded stage.
You apply
Four short steps against one specific role. When you submit, you get a reference number you can use to check the recorded stage. We also email it when delivery succeeds.
Your application is assessed
C.O may compare what you sent with the role and prepare a read for our hiring team. Submitting an application does not guarantee a personal read.
First discussion
If your application moves forward, we may invite you to discuss what you've built, what you want to build, and the work we're doing.
Working session
For some roles, the next step is a working session on a problem like one you'd meet here. We won't keep unpaid project work.
Check the outcome
Your reference shows the latest recorded stage. We email you when that stage changes, including when we close an application.
See who it's actually for.
Before you decide this is worth building, read what it's changed for the people who use it, work with it, or work here — published as they wrote it, not as we'd have put it.