Jobs · Tracker · Resume · Coach

From a job listing to an offer, in one continuous thread.

Other apps stop at "here are some jobs." Dilly takes you the rest of the way. The feed surfaces roles that fit your field. Every listing gets a per-job narrative explaining whether you're Ready, Almost, or Gap. The tracker logs every move automatically. Forge writes a tailored resume from real profile facts. And Dilly the AI remembers all of it the next time you open the app.

Get Dilly free → See the flow

1The feed

A jobs feed tuned to your field, not the universe.

Pick your field and your mode once. Dilly's feed pulls from 30+ ATS sources and refreshes nightly. You see roles built for who you are right now: a Data Science student looking for a summer internship gets a feed of internships, not VP roles.

50,000+ active listings updated nightly across 49 fields
Mode-aware. Student favors internships and entry-level; Seeker shows full-time roles
Rejections don't reappear. Turn a role down once and the family is downranked in tomorrow's feed

Jobs feed

Data Science Intern● Ready

Roblox · San Mateo, CA · Summer 2026

ML Engineer, New Grad● Almost

Anthropic · SF / Remote · Full-time

Quantitative Researcher● Gap

Citadel · NYC · Senior level

Fit narrative · Roblox

Data Science Intern● Ready

Roblox · San Mateo, CA

Why Ready: Your KVR project shipped a SQL pipeline serving 60+ active members. Roblox specifically calls out "ad-hoc SQL on production data." Your Snowflake cert closes their declared stack gap. Tonight's move: Apply, and ask Dilly to draft a 3-bullet pitch from your profile.

2Fit narrative

Not a score. A narrative that explains itself.

Every listing gets a paragraph, not a number. Ready means you clear the bar. Almost means one gap. Gap means the role is beyond your current state, and the narrative says exactly why.

Sourced, not invented. Each "why" line maps to a real profile field
Specific, not generic. "Your KVR project", not "your past experience"
Actionable. Every narrative ends with one or two moves you can make tonight

3Forge

One tap. A resume built for that exact role.

From the job card, tap Tailor in Forge. Dilly drafts a resume for that specific listing, drawn entirely from facts already in your profile. The Truth Ledger blocks anything the AI tries to invent. You export a PDF; the application uses real wording you can defend in the interview.

100% sourced. Every bullet points to a profile field, visible in the editor as a chip beneath the line
Per-role variants. The Roblox draft reads differently from the Anthropic one
No hallucinated achievements. If a claim has no source, it does not ship

Forge · Roblox draft

Led the KVR data science club through weekly ML workshops, growing membership 60% in one semester. Source: experiences.president_data_science_club

Built a Snowflake-backed SQL pipeline serving 60+ members; sub-second queries on 4M rows. Source: projects.kvr_pipeline + certifications.snowflake

Shipped a user-facing engagement dashboard tracking weekly retention and chapter activity. Source: projects.kvr_dashboard

Your tracker

Roblox · DS InternAppliedTonight

Anthropic · ML EngPhone screenFriday

Microsoft · SWE InternRejectedTuesday

Datadog · APMOfferLast week

4The tracker

Every application logs itself. The feed gets sharper.

The moment you tap Apply, the tracker records it, role, company, date, the version of your resume that went out. When you log a rejection or an offer, the AI uses it. Get rejected from a specific role family three times, and similar listings get downranked in tomorrow's feed automatically.

Auto-logged. Every apply, screen, offer, and rejection writes to the same record
Feeds the feed. Rejections reweight the next round; offers tune fit narrative confidence
Visible to chat. Ask Dilly "what's next?" and it sees the whole pipeline before it answers

5The AI Coach

One AI tying the four together. Remembers everything.

The feed, the narratives, the resume, the tracker, all read from the same profile and surface through the same AI. Open chat at 11pm after a rejection, and Dilly already knows the recruiter feedback, the gap they flagged, the next three roles that fit better, and the mock interview already on your calendar.

Already loaded. Profile, tracker state, last seven days of activity, all in context before your first message
Writes facts back. Tell Dilly something true and the profile saves it
Suggests the next move. Specific actions tied to specific listings or specific gaps

Chat · Tuesday 11:47pm

Got rejected from Microsoft again. I'm kind of done.

I know. The recruiter note came in at 4pm and it's the same gap they flagged in February. I held three things back so you didn't have to think about them tonight: two Series B roles that fit better than Microsoft did, a one-week sprint of 4 Skills chapters that closes the gap, and a mock Saturday 10am with the engineer who prepped you for Datadog. All on your calendar.

Why this works

Four surfaces, one brain, one profile.

The feed, the resume, the tracker, and chat all read from the same profile and write back to it. That's why a rejection on one screen sharpens the recommendations on every other screen tomorrow.

Reads from profileJobs feedField, mode, recent activity. Outputs a tuned list with per-role fit narratives.

The brainDilly AI · ProfileMaster AI reading everything, writing durable facts back. The reason every screen knows what every other screen did.

Writes backTrackerApplies, screens, offers, rejections. Every event becomes a signal that retunes tomorrow.

Reads from profileForge resumeTailored to the listing, sourced to your profile, blocked from inventing anything.

Closes the loopAI Coach chatReads the whole pipeline. Writes durable facts back. Suggests the next concrete move.

Reads from profileFit narrativesPer-job paragraphs. Each "why" line maps to a real field. Each ends with a move.