Blueprint · Hiring & Mobility · Autonomy Tier: 1

Turn fuzzy headcount goals into a concrete, tech-ready sourcing plan.

This blueprint defines how Sourcing turns a vague “we need more engineers” request into structured, repeatable sourcing strategy for engineering, data, and AI roles — including Boolean patterns, target companies, and market signals — while keeping recruiters in control.

← Back to Home

Tools System Cases Articles

View Tool Overview What this tool solves, outcomes, and signals. View Tool Demo Try the interactive tool.

What this Blueprint is

Problem, mission, and outcomes for tech-native sourcing · scope, guardrails, and design principles · inputs, signals, and role templates · Boolean & X-ray patterns (LinkedIn, GitHub, open web) · company & market mapping for tech ecosystems · calibration, governance, and Metrics outputs.

Who this Blueprint is for

TA leaders, recruiting ops, and tech recruiters who need structured sourcing patterns for engineering, data, and AI roles — without turning the process into a black box.

What this tool solves

Vague briefs → blank search bars | Rebuilding Boolean from scratch | Personal bookmarks vs shared patterns | Scattered company mapping | Hard-to-debug top-of-funnel dryness

1. Problem & Mission

1.1 The problem

Most tech sourcing starts with a vague request and a blank search bar:

Hiring managers say “we need great engineers” without a clear market view.
Recruiters rebuild Boolean and target lists from scratch every time.
Sourcing quality depends on personal bookmarks, not shared patterns.
Company mapping lives in scattered sheets and memory.
Effort is hard to measure and harder to debug when pipelines dry up.

1.2 Mission

Provide a structured starting point — market map, target companies, and Boolean patterns — in minutes instead of days, while keeping recruiter judgement in the loop.

1.3 Outcomes

Reusable sourcing blueprints for common tech roles.
Faster time-to-first-outreach when a role opens.
Visibility into which markets and companies are targeted — and why.
Structured sourcing outputs that feed downstream tools and measurement.

2. Scope & Design Principles

2.1 In-scope

Engineering, data, ML/AI, DevOps, and security roles (v1 focus).
JD + brief → structured role profile.
Boolean + X-ray patterns for LinkedIn, GitHub, and open web search.
Company mapping: competitors, adjacencies, “hidden gem” sources.
Market sizing estimate (small / medium / large pool).
Structured outputs that plug into screening + measurement.

2.2 Out-of-scope (v1)

Automated outreach, nurture, or sequencing.
Replacing recruiter judgement on who to contact.
Real-time scraping or user-level behavioral tracking.
Predicting response rates beyond simple assumptions.

2.3 Principles

Pattern-first, tools-second. Logic should survive tool changes.
Explainable search logic. Every pattern is readable and editable.
Tech-ecosystem aware. Roles are defined in real stacks and company types.
Human-in-the-loop. Recruiters approve, refine, and save patterns.
Non-creepy by default. Public professional signals + market insights.

2.4 Global rules

No automated candidate outreach without recruiter review.
No inference or storage of protected characteristics.
Company mapping can be shared; candidate data stays controlled.
Templates and patterns are versioned and discoverable.

3. Roles & Responsibilities

Role	Responsibilities
Tech Recruiter / Sourcer	Inputs hiring brief, edits patterns, runs sprints, saves winning templates.
Hiring Manager	Aligns on profiles, targets, and stacks; reviews market map and signs off.
Recruiting Ops	Owns templates, taxonomies, and integrations; drives consistency.
People Analytics	Analyzes top-of-funnel health, market coverage, and sourcing ROI.
Owner	Diane Wilkinson — sourcing logic, ecosystem mapping, continuous improvement.

Key idea: Sourcing does the heavy thinking on where/how to search; humans decide who to approach and how to tell the story.

4. System Overview

1. Inputs: JD, brief, stack, location, seniority, must-haves vs nice-to-haves.
2. Role profile: Convert requirements into a structured tech role template.
3. Market mapping: Target companies + adjacent pools.
4. Search strategy: Boolean + X-ray strings (LinkedIn, GitHub, web).
5. Readiness: Rough pool sizing + recommended channel mix.
6. Output: Save profile + patterns + company list as reusable artifacts.

4.1 Technical architecture

Parsing layer (LLM + rules) to extract attributes from JD/brief.
Role template library (backend, ML, data, DevOps, security).
Pattern engine to assemble Boolean / X-ray / GitHub searches.
Optional company index (CSV/DB) for tiering and mapping.
Light integrations to ATS/CRM fields or internal sheets.

5. Inputs & Signal Library

5.1 Core inputs

Field	Description
role_title	Internal label (e.g., “Staff ML Engineer”).
role_family	Engineering, Data, ML/AI, DevOps, Security.
seniority_level	IC / Senior / Staff / Principal / Manager.
primary_stack	Languages/frameworks (Python, Go, React, K8s, PyTorch).
must_have_skills	Non-negotiables.
nice_to_have_skills	Preferences that expand pool.
target_locations	Regions / remote constraints.
company_archetypes	Ideal employer types/verticals.
avoid_list	No-go companies/types.

5.2 Signals used in patterns

Title synonyms: “ML Engineer”, “Applied Scientist”, etc.
Stack synonyms: “MLOps” variants, “LLM” variants.
Location logic: city/metro/time-zone remote.
Company tiers: competitors, adjacent infra, hidden gems.
Channel fit: LinkedIn vs GitHub vs niche communities.

Note: This is about “where to look and what to look for” — not guessing culture fit.

6. Search Strategy & Boolean Patterns

6.1 LinkedIn Boolean — overview

Title cluster
Skill/stack cluster
Domain keywords
Negative filters

6.2 Example — Senior Backend Engineer (Python, distributed systems)


          ("backend engineer" OR "back-end engineer" OR "software engineer")

          AND ("Python" OR "Go" OR "Golang")

          AND ("distributed systems" OR "microservices" OR "high scale" OR "high-throughput")

          NOT ("intern" OR "student" OR "bootcamp")

6.3 GitHub search — overview

Language filters
Repo keywords
Engagement signals (stars, forks, contributions)

6.4 Example — ML Engineer (LLMs)


          ("LLM" OR "large language model" OR "transformers")

          ("PyTorch" OR "TensorFlow" OR "JAX")

          language:Python

          stars:>10

6.5 Google X-ray (LinkedIn)


          site:linkedin.com/in

          ("Senior Machine Learning Engineer" OR "Applied Scientist" OR "ML Engineer")

          ("LLM" OR "large language model" OR "RLHF" OR "transformer")

          ("San Francisco" OR "Bay Area" OR "San Jose" OR "Oakland")

          -"student" -"intern"

7. Company & Market Mapping

7.1 Company tiers

Tier	Description	Examples (AI/ML)
Tier 1 — Direct competitors	Core labs and direct problem competitors.	Anthropic, OpenAI, Google DeepMind, Cohere, xAI
Tier 2 — Adjacent & infra	Applied AI, infra, tooling with overlapping skills.	Scale AI, Databricks, Snowflake, Hugging Face, W&B
Tier 3 — Hidden gems	Mid-stage startups, OSS-heavy teams, niche verticals.	Series B–D AI startups; OSS-heavy fintech/health/devtools

7.2 Company index fields

company_name, parent_company, group
tier (1/2/3), vertical, stack_tags

8. Calibration & Governance

Run patterns in shadow mode alongside manual patterns.
Track which patterns produce qualified responses and screens.
Promote winners into a shared template library.
Version templates; restrict core taxonomy edits to Ops.

9. Integration & Handoffs

Artifact / Field	Used by	Purpose
role_profile_id	All tools	Connects sourcing → screening → interview workflow.
sourcing_boolean_linkedin	Recruiters	Primary LinkedIn pattern for sprints.
sourcing_boolean_xray	Recruiters	Google X-ray pattern for public profiles.
sourcing_pattern_github	Recruiters	GitHub pattern for engineering roles.
target_companies_tier1_3	Recruiters & HMs	Shared company focus list with rationale.
estimated_pool_size	Measurement	Pipeline planning and expectations.
sourcing_channel_mix	Measurement	Planned vs actual channel mix.

10. Metrics & Measurement

10.1 Key sourcing metrics

Prospects contacted per week (by role, recruiter, channel).
Warm response rate (reply/interest) by channel and pattern.
Prospect → screened conversion by role template.
Screened → onsite/offer conversion by sourcing pattern.
Hires by company tier and sourcing channel.

Over time, you can say: “For Staff ML roles, our best hires came from Tier 2 infra using Pattern B — not Tier 1 labs using Pattern A.”

Appendix A — Example Tech Role Profiles

A.1 Senior Backend (Python / distributed systems)

Titles: Senior Software Engineer, Senior Backend Engineer
Stack: Python/Go, microservices, cloud, distributed systems
Companies: SaaS, infra tooling, devtools, high-scale apps
Channels: LinkedIn, GitHub, engineering communities

A.2 Staff ML (LLMs)

Titles: Staff ML Engineer, Applied Scientist, Research Engineer
Stack: Python, PyTorch/JAX, LLMs, MLOps, GPU optimization
Companies: labs, applied AI, infra/tooling, OSS-heavy teams
Channels: LinkedIn, GitHub, ML communities

Appendix B — Example Boolean & X-ray Strings

B.1 Staff ML — LinkedIn Boolean


          ("staff machine learning engineer" OR "staff ml engineer" OR "applied scientist" OR "research engineer")

          AND ("LLM" OR "large language model" OR "generative AI" OR "transformer")

          AND ("Python" OR "PyTorch" OR "TensorFlow")

          AND ("deployment" OR "production" OR "serving" OR "MLOps")

          NOT ("intern" OR "student" OR "bootcamp")

B.2 Backend — X-ray


          site:linkedin.com/in

          ("Senior Backend Engineer" OR "Senior Software Engineer")

          ("Python" OR "Go" OR "Golang")

          ("microservices" OR "distributed systems")

          ("San Francisco" OR "Bay Area" OR "Remote")

          -"student" -"intern"

Let's Connect

Open to roles in People Analytics, Talent Intelligence, People Ops, and Recruiting Operations — especially teams building internal AI capabilities.

📩 Email Diane 🔗 Connect on LinkedIn 📅 Schedule 15 min Intro