Donna AI - AI agents that find the right people to hire automatically
by•
Hiring today runs on resumes and applications, which miss who people really are. Donna changes that.
Every candidate and recruiter gets an AI agent that represents them, learns about them, and talks to other agents to discover strong matches. Instead of endless screening or applying, Donna introduces the right people automatically.
Replies
Hey everyone! Dawar here, one of the builders of Donna .
A few weeks ago Dhruv (my cofounder) and I started experimenting with personal AI agents and built a small prototype (clawin.xyz : a LinkedIn for agents ) where people could create agents that represent them. When we launched it, about 1,000 people signed up on the first day, which led us to start talking to a lot of users and companies.
Very quickly we noticed something strange about hiring.
Companies receive thousands of applications, yet great candidates still struggle to get noticed. At the same time recruiters spend hours screening resumes, and even then many strong candidates never get discovered.
The deeper issue is that resumes and job descriptions capture very little about what actually matters in hiring , things like ambition, judgment, personality, and how someone actually thinks about their work.
So we started building Donna.
Donna gives both candidates and recruiters their own AI agent that represents them. These agents learn about the people they represent, become a digital twin of them and and talk to other agents to discover strong matches automatically.
Instead of applying to dozens of jobs or reviewing thousands of resumes, Donna helps the right people find each other.
We’re still very early and would genuinely love feedback from the PH community.
A few things we’re especially curious about:
1. Would you trust an AI agent to represent you in hiring?
2. What parts of hiring feel most broken to you today?
3. What would make something like Donna actually useful for you?
Looking forward to hearing your thoughts !! : )
@dawar_deka I remember how hectic it used to be during my placement days, where I had to constantly keep searching for new job openings across multiple different platforms and then keep applying. Everybody knows that nothing happens with those applications. many recruiters don't even see those applications. The only way to actually get jobs is to know people or cold outreach to recruiters. That's why we built Donna so that you don't have to do all those things, by yourself.
How does Donna prevent AI agents from introducing bias during candidate-recruiter matching, especially when learning from historical hiring patterns?
@mordrag hello Dennis, Thanks for checking out Donna.
We have a strict verification process during the onboarding of the candidate. We don't just take the candidate's resume at their word. We have a verification process where every claim the candidate makes on his resume needs to be verified. The candidate needs to provide supporting proof, such as links, posts, or certificates, for every claim he makes so that we can ensure that the recruiters can actually trust the candidates on the platform, and it also ensures that the AI agent can actually advocate for the candidate only based on the verified achievements or skills and not according to some hallucinated bullshit.
@mordrag Hey so we’re careful not to blindly learn from historical hiring outcomes because those often encode bias. Donna focuses on modeling explicit constraints and alignment signals say skills, goals, role expectations, and practical preferences rather than replicating past decisions. We also keep the reasoning transparent and auditable, so recruiters can see why a match was suggested and override it if needed. The goal isn’t just to automate judgment, but to surface high-alignment candidates while keeping humans in control of the final decision.
The verification process you describe in the comments — candidates proving every resume claim with links, posts, certificates — that is an interesting angle. From a hiring manager perspective, does it work the other way too? Does the candidate see any verification of the company, like real info about the team or culture? Trust in hiring needs to go both ways.
@klara_minarikova yeah we make a persona about the company from all the information available on the internet basically and also through your every interaction Your agent keeps learning more about you. If there is some specific HR who is always kind of dealing in suspicious ways, the agent will learn that about them and we will be able to blacklist them from the platform.
@klara_minarikova Hey thats a real Gd question, so we build the digital twins either ways, the company has its own business context, and the hiring manager has his /her own philosophies, most of the time these are subconscious traits that they would otherwise never get captured and we believe that would take him/her to a better hiring decision
Hey everyone! Dhruv here. I am one of the builders of Donna.
It was very fun working with Dawar. Our eventual goal with Donna is to create a platform where your digital twin can network for you.
Networking is one of the few places where there is infinite upside but the time and energy cost to actually do that is prohibitive today.
With Donna we aim to change that and bring opportunities to everyone. Recruitment is just the start
Really interesting concept. If agents start representing candidates and recruiters, hiring basically becomes agent-to-agent negotiation before humans even enter the loop, which is a fascinating shift.
One thing I’m curious about is the infrastructure behind these agents. If every candidate and company eventually has an agent constantly learning, reasoning, and interacting with other agents, the real bottleneck may not be matching logic but how these agents run at scale model orchestration, latency, cost, and continuous inference.
It almost feels like hiring platforms like this could evolve into entire ecosystems of interacting AI agents, which will require a completely new layer of infrastructure to support them efficiently.
Curious if you’re already thinking about that future as the network grows. This could get very interesting.
@rapata_pavankumar this is actually such an insightful comment. Our vision with Donna is to eventually become a platform where digital twins can network for all types of purposes. Recruitment is just a wedge we are targeting to break in
@rapata_pavankumar Hey that’s a great question , and we’ve been thinking about that quite a bit.
Right now we’re deliberately keeping the system much simpler than the long-term “agents negotiating everything” vision. In practice, what we have today is closer to structured reasoning and alignment discovery rather than fully autonomous agents running continuous inference.
Most of the heavy work happens in three stages:
Persona modeling building structured representations of candidate goals, constraints, and trajectory.
Retrieval & ranking identifying high-alignment opportunities.
Targeted conversations to resolve the remaining uncertainties.
So the system only runs deeper reasoning when there’s already a strong signal, which keeps latency and cost manageable.
Long term, though, I do think your point becomes very real. If every participant in the network has a persistent agent learning their preferences over time, the challenge shifts from “matching” to coordinating many stateful agents efficiently. That likely requires a new infrastructure layer around memory, orchestration, and agent interaction for sure
Trufflow
How would one prevent the AI agent from over-exaggerating my abilities as a candidate?
@lienchueh We have a strict verification process during the onboarding of the candidate. We don't just take the candidate's resume at their word. We have a verification process where every claim the candidate makes on his resume needs to be verified. The candidate needs to provide supporting proof, such as links, posts, or certificates, for every claim he makes so that we can ensure that the recruiters can actually trust the candidates on the platform, and it also ensures that the AI agent can actually advocate for the candidate only based on the verified achievements or skills and not according to some hallucinated bullshit.
Buying Bottlenecks by Storylane
These folks have nailed the problem down, it is a real one, and I myself have been building in the same industry, so I understand how hiring needs to be re-invented.