Navia Logo

Decoding the Admissions BlackBox with AI

Portrait of Amir Rassouli

Amir Rassouli

Co-Founder

Few experiences test human patience more torturous than waiting for an answer you have no control over.
The refresh of an inbox, the silence after months of work, the creeping sense that the application process is rigged by factors you’ll never see. Everyone tells you that admissions are about merit, but anyone who’s been through it knows
it’s also about timing, invisible rules, and accidents of luck.

Act I: The Signal-to-Noise Ratio Problem

As international graduate application volume explodes and acceptance rates contract, the core challenge for any applicant is to emit a signal strong enough to cut through the noise.
This visualization shows the growing disparity.

Act II: The Information Assymetry Problem

Admissions committees make their decisions using real-time and insider data (things like funding availability and faculty hiring needs) that applicants never get to see. Meanwhile, applicants have to make choices using a partial and outdated picture of the situation.

Act III: The Strategic Miscalculation Problem

The direct result of this environment is a critical misallocation of effort. Without clear data, applicants overweight factors they can control (like test scores) and underweight factors that committees secretly prioritize (like specific research alignment). This disconnect is where most applications fail before they're even read.

How about we reverse engineer that?

ADMITTED COHORT Research Alignment Quantitative Scores Collaboration Affinity Outreach Timing YOUR PLAYBOOK analyze build

We turn the question “why was this person admitted?” into a reconstructive framework: deriving explanatory patterns from observed outcomes, underlying signals, and contextual constraints, and then operationalizing them for planning. The goal is to develop an AI agent that systematically narrows the gap between applicant-controllable factors (like cross-domain positioning, timing, narrative alignment) and non-controllable ones (like institutional preferences, cohort dynamics). This agent reframes the application process from a speculative venture into a data-driven optimization task, and actively reduces uncertainty and risk in decision outcomes. What follows illustrates how it exactly does that.

Answer I: You’re Lost in the Noise, So You Need a Direction

Most applications shout into static.
Our model maps your position among real cohorts, quantifies your signal, and traces routes that measurably raise admit probability; so every step is intentional, not hopeful.

Answer II: You’re an Idea in Motion, So You Need the Right Orbit
(Case Study For an Iranian Applicant)

Our model quantifies alignment (research overlap) and affinity (collaboration patterns with students from your background and region) across the supervisor market.

This is an illustrative example of our product, not actual data.

Answer III: You're a Candidate, So You Need a Campaign

Opportunity moves fast.
Our model designs a structured, three-wave supervisor outreach sequence that balances personalization depth with expected ROI and keeps momentum until meetings land.

The Ladder of Agency

We're building the world's first fully agentic application machine. Today, we've achieved critical milestones. Here's where we are now, and the path ahead.

Available

L1

Insight-level Agency

From "What?" to "What If?"

Analyzes your CV to map your position, quantify strengths against real cohorts, model admission probability, and trace the most efficient improvement pathways.

Available

L2

Strategic-level Agency

From "Where?" to "Who and How?"

Scans the entire academic landscape to find supervisors whose needs align with your profile, then structures a multi-wave outreach plan based on cost-benefit analysis.

Almost Available

L3

Execution-level Agency

From "Plan" to "Progress"

Delivers hyper-personalized outreach at scale --- blending audience insights with recent activity and continuously optimizing performance in real time.

It is coming

L4

Autonomous-level Agency

From "Campaign" to "Admission"

Provides comprehensive, end-to-end execution --- turning your objectives into results through autonomous planning, action, and optimization.

Only your CV. That’s all it takes.

The process begins with a raw CV or Resume, from which the model performs an initial diagnostic assessment to determine the candidate’s current positioning (“where you stand”). It then autonomously infers a progression path (“where to move”) and defines strategic objectives (“what to target”), finally operationalizes these insights through data-driven precision outreach to achieve measurable outcomes. You can try it out here.

Acknowledgement

I’m enormously grateful to my co-founders: to Komeil, for sacrificing his sleep, sanity, and weekends in pursuit of perfection; and to Mahdi, for disappearing to build his so-called “agents“ that might one day replace us as founders. We can only hope one of those agents will eventually handle his communications with us, since he won’t. Without both of them, none of this beautiful chaos (or success) would’ve been possible.

My deep appreciation goes to our Head of Marketing, Hamed, whose presence has been simply irreplaceable. I’d also like to thank my dear friend Mojtaba for his invaluable and genuine support, especially during those dark days.