NetBramha Recruitment Portal

NetBramha Studios · Internal Tool Suite · PM Series 09–10

Recruitment Portal

A connected stack of two tools, four platform shortcuts, and one shared data layer — covering the full hiring journey from sourcing a profile to moving a candidate through the pipeline. Built for Raghav and Muskaan. No third-party ATS. No context lost between steps.

Live · Internal

Built by Sarath MS
@netbramha.com gated
Claude Sonnet · Google OAuth
2026

The Stack — How the Tools Connect

Sourcing Shortcuts

LinkedIn · Weekday · Zoho · Naukri

Browser shortcuts that run on any platform profile page. Extract candidate data and pre-fill the Add Candidate form in one click — no copy-pasting.

4 platforms Claude in Chrome Bookmarklets

›

First Cut

netbramha-shortlist.netlify.app

AI resume screener. Reads inbound applications from Gmail or batch-uploaded PDFs. Scores each resume against the open role. Pushes shortlisted candidates to the pipeline.

Mailbox Batch Upload AI Scoring Push to HI

›

Hiring Intelligence

netbramha-hiring.netlify.app

Pipeline dashboard. Tracks every candidate across all active roles. Stage changes, history, notes, and AI summaries all on one screen. One sheet, one source of truth.

Kanban Role View Talent Pool AI Forecast

Single shared data layer → Google Sheet · 1smuoTDeziyd... · Daily Tracker + Roles + Claims tabs · Worker: hiring-worker.rey619-ms.workers.dev

Sourcing Intelligence — Outbound Sourcing Workflow

Set up Scoring Params

Paste the JD into Sourcing Intelligence → Scoring Params tab → Generate Parameters. AI extracts weighted criteria (must total 100). Save per role.

›

Generate Search String

Tool generates an optimised Naukri/LinkedIn search string from the JD and saved params. Copy and paste directly into the platform search bar.

›

Generate Shortcut

⚡ Generate Shortcut creates a role-specific Claude in Chrome command (e.g. /screen-bde). Save it once. It embeds your scoring criteria.

›

Screen on Naukri

On the Naukri listing page (40 cards loaded), type /screen-bde. Claude reads all cards, scores against embedded criteria, and ticks the top candidates. No profile opens needed.

›

Bookmarklet Sends + Ranks

Click NB Sourcing bookmarklet. Reads ticked cards, extracts structured data, sends to Worker. Worker AI-scores against saved params → writes ranked results to sheet. Sourcing Intelligence opens automatically.

Ranked Results — What You See

#1

Candidate Name

Current Role · Company · 4 yrs

87

/100

#2

Candidate Name

Current Role · Company · 2 yrs

71

/100

#3

Candidate Name

Current Role · Company · 5 yrs

58

/100

Load from Sheet → ranked by AI score. Each card shows matched signals and red flags. Push to HI when ready to engage.

Naukri

Primary outbound platform. Listing-page bookmarklet reads all ticked cards (up to 40) in one click. Shortcut pre-screens before bookmarklet fires — only quality candidates land in the sheet.

Weekday

Profile-level bookmarklet. Click on any candidate profile — extracts name, LinkedIn URL, role, company, and pushes directly to HI via upsert. Source: Weekday · Direction: Outbound.

Zoho

Reads EntityInfo.do pages. Extracts name, email, phone, associated role. Pushes directly to HI. Source: Zoho Applications · Direction: Inbound.

LinkedIn

Profile bookmarklet reads h1 name, headline, Education, and Experience sections. Push to HI. Or use /push-linkedin shortcut for richer extraction with a pre-filled confirmation step.

Other sources tracked in HI →

Careers Site · Inbound Employee Referral Campus MetaSkills / Vendor Recruz

Shortcut expires after 55 min → re-activate from Sourcing Intelligence → Activate Shortcut

First Cut — Inbound Resume Screening

Mailbox Mode

Connect the inbox

Connect careers@netbramha.com via Google OAuth

First Cut reads unread emails with resume attachments

Click an email → select role → run AI Review

Review scores, edit summary, check candidate fields

Push to Pipeline or send rejection email in one click

Batch Upload Mode

Drop up to 20 resumes

Select the open role from the dropdown

Drag and drop PDF or Word files — up to 20 at once

AI scores each resume in parallel — Experience + Fit, both 0–10

Review cards. Red flags and 3-line hiring brief per candidate

Push individual cards or all shortlisted in bulk

Strong Yes

Total ≥ 15 · no critical criteria missing → Auto-staged as Shortlisted

Maybe

Score 8–14 · genuine uncertainty → Staged as On Hold · human call needed

Reject

Total ≤ 7 or must-have absent → Rejection email sent via Gmail API

"Candidate has 3 years listed but graduated 2 years ago — timeline doesn't add up. Visual design background, no evidence of UX research ownership. Portfolio shows UI only."

That's the AI brief. Not a template. Specific to the candidate, specific to the JD. Editable before pushing.

Hiring Intelligence — Pipeline Stages

1 → 2

Reached Out → Shortlisted

Outbound candidate contacted or inbound application cleared First Cut. Enters active pipeline.

Task Sent

Assessment or design task sent. Waiting on submission. HI tracks the date automatically.

Interview Scheduled

R1 · R2 · F2F · Mock Call · G-Meet. All map to this stage with notes for the specific round.

Offer / Joined

Offer extended or candidate has joined. Closes the loop on the role card metrics.

8 · 9

Pipeline · On Hold

Good candidate, wrong timing or role. Tagged for Talent Pool re-engagement when a matching role opens.

Hiring Intelligence — Key Features

🗂

Command Centre

All active roles ranked by priority (P1 → P3). Each role card shows open count, pipeline count, days open, and stage distribution. One view to understand where hiring effort needs to go.

📋

Kanban per Role

Drill into any role for a full stage-by-stage kanban. Click any candidate card to open their history — all stage changes with timestamps, notes, AI score, and recruiter rating.

🔒

Multi-User Claim Locking

When Raghav opens an email in First Cut, it soft-locks for 30 minutes so Muskaan doesn't review the same candidate simultaneously. Claims tab in the sheet tracks ownership.

♻️

Talent Pool

Re-engage tab surfaces silver-medal candidates — strong profiles rejected for timing, location, or role mismatch. Priority-scored against new openings. Outreach templates included.

⚠️

Repeat Applicant Detection

First Cut flags candidates who have applied before. An amber chip on the inbox card shows their prior stage, role, and date — before the recruiter opens the email.

🤖

AI Role Forecast

Claude analyses pipeline data for a role and forecasts hire likelihood, likely bottleneck stage, and recommended actions based on historical stage conversion patterns.

The Data Layer — One Sheet, Everything

Role

D · E

Name · LinkedIn

F · G

Direction · Source

J · K

Stage · Notes

History (log)

N · O · P

AI Score · Rating · Verdict

Neither tool manages the sheet directly. All reads and writes go through the Cloudflare Worker → /candidate/upsert deduplicates automatically. Daily Tracker + Roles tab + Claims tab + FC Roles tab.

Stack

Netlify · Single HTML files Cloudflare Worker · hiring-worker.rey619-ms.workers.dev Claude Sonnet · Vision + Text scoring Google Sheets API · shared data layer Gmail API · OAuth read + send Google OAuth · @netbramha.com gated Claude in Chrome · platform shortcuts PDF.js · OCR fallback for scanned resumes mammoth.js · Word file support No ATS. No build step. No external vendor.