How Tablah Works

Not a CV generator. A career intelligence system that runs your job search in the background and builds your application materials from verified data.

Ready for a Guided Setup?

Follow our step-by-step checklist to get your Vault populated, your feed running, and your first application workspace open. Takes about ten minutes to get your first job matches.

Tablah system architecture flow

1. The Experience Vault

The Vault is a project-based career repository β€” not a resume. Every experience is parsed into structured entries across four dimensions: Skill, Domain, Seniority, and Behavior. All AI-generated materials are grounded strictly in what you put here.

πŸ”

Three Ways to Import

Upload a CV (PDF or DOCX), paste raw text, or use the Chrome extension to pull your history directly from LinkedIn. The AI parses and structures the entries β€” you review and confirm.

⚑️

Project-Based, Not Job-Title-Based

Each entry captures the tasks, tools, stakeholders, and measurable outcomes of a specific role or project. The AI tags every entry across Skill, Domain, Seniority, and Behavior facets.

πŸ”„

Representative Experiences & Career Centroid

Flag the roles that reflect where you are now. The system calculates your Career Centroid: a 3072-dimension vector embedding that powers your job feed and CV generation. Older, less relevant roles can stay in the Vault without diluting your match scores.

βœ…

Attestation & Integrity Badges

Former managers can sign off on specific entries directly in the platform. A cryptographic hash of the text is stored β€” if you later change the wording to exaggerate your role, the badge is automatically downgraded or revoked.

2. The Discovery Feed

The Feed runs a global job sync daily and scores every result against your Career Centroid using a three-stage AI pipeline. No manual keyword searching β€” the system surfaces what fits, ranked by how well each role aligns with your actual profile.

🎯

Semantic Matching, Not Keywords

Your Career Centroid is compared against every job in a 3072-dimensional space using vector search. The algorithm understands contextual meaning β€” 'Node.js' and 'server-side JavaScript' are treated as equivalent, not as a mismatch.

πŸ“ˆ

Three-Stage Scoring Pipeline

Stage 1: fast ANN retrieval of the top ~500 jobs. Stage 2: re-ranked by your personal Match Tuning weights across the four facets. Stage 3: a Gemini LLM deep-dives the top 10–20 candidates for a Diamond Fit score with a full radar breakdown.

πŸ”

Match Tuning & Fit Radar

Adjust how much weight the algorithm gives to Skill, Domain, Seniority, and Behavior. The Diamond Fit radar shows exactly where you match and where the gaps are β€” so a high score is explainable, not a black box.

πŸ”€

Experience vs. Career Goal View

Toggle the feed between your current experience profile (backward-looking) and a future-facing Career Goal view β€” useful if you are pivoting into a new role or industry and want matches weighted towards where you are heading, not just where you have been.

3. The Job Pipeline

Your private shortlist of jobs worth pursuing. Save from the feed, import from any job board by URL, or capture via the Chrome extension. Every job gets the same scoring treatment regardless of how it arrives.

πŸ“Œ

Import from Anywhere

Paste a URL from LinkedIn, Indeed, or any other board and the AI scores it against your profile. The Chrome extension lets you import with one click while browsing.

🌍

JD Analysis & Company Research

Before you apply, the AI scans the job description for red flags, contradictions, and inflated requirements. It also surfaces company research so you know what to probe in an interview.

πŸ”

Kanban Status Tracking

Track every application from Saved through Interviewing to Offer in a single visual board. The full history of your applications, communications, and generated documents stays attached to each card.

4. The Application Workspace

Per-job tailoring factory. Once you decide to apply, the workspace builds your materials from the Vault β€” the Action Plan identifies gaps and strengths, the CV engine drafts and iterates, and the Proof-reader catches anything that cannot be backed up.

πŸ“‹

Action Plan: Mind the Gap

The AI runs a gap analysis between your Vault and the job description. Study items flag genuine knowledge gaps with a dedicated research chat. Highlight items surface skills you already have but haven't made visible enough in your narrative.

✍️

ATS-Optimized CV Generation

The engine selects your 3–5 most relevant experiences and drafts a CV in your chosen format. Missing keywords are injected via the Skill Vault Bridge β€” grounded in a real experience you select, not invented. 'Learn My Style' applies your editing preferences to future drafts.

πŸ—ΊοΈ

AI Critique (Proof-reader)

Before you submit, the Proof-reader scans the tailored CV against your raw Vault data and flags any claim it cannot substantiate. One-click fix re-grounds the language in verified facts β€” so you are not caught off-guard in an interview.

🎀

Interview Prep & AI Representative

Generate likely interview questions with vault-backed suggested answers. Practice with an interactive mock interviewer. Share a secure link that lets a recruiter-facing AI chatbot answer questions about your profile 24/7 β€” grounded strictly in your verified experience.

Ready for a Guided Setup?

Follow our step-by-step checklist to get your Vault populated, your feed running, and your first application workspace open. Takes about ten minutes to get your first job matches.

* The Diamond Fit AI reranker, Match Map, and Attestation engine require a Pro subscription.

Start Your Pipeline Now