Software engineer job search playbook: discovery → tailor → apply

This playbook is for software engineers who want more interviews per hour spent — not more applications per hour. It covers discovery → match → tailor → apply, with automation where it helps and human judgment where it matters.

Phase 0: One-time profile setup (2–4 hours)

Before searching, build a complete job search profile:

• Resume + extended project writeups
• Preferences: remote, seniority, visa, dealbreakers
• Behavioral stories for form questions
• Target narrative: what you want next and why

This upfront investment pays back on every subsequent role.

Exit criteria: You could paste accurate answers into a Greenhouse form without opening old docs.

Phase 1: Discovery (ongoing, automated)

Manual discovery does not scale. Effective channels for tech:

Channel	Role
Lever / Greenhouse / Ashby	Primary — structured URLs and forms
Referrals / network	Highest conversion — still do this
LinkedIn	Supplement — noisy but useful for signals
Aggregators	Discovery only — apply on company ATS when possible

PrismApply runs curated Google searches (Serper) against ATS domains, dedups known URLs, scrapes new postings, and embeds job content for matching.

Your job: Set discovery-friendly preferences so matches are actionable.

Phase 2: Match (automated + gates)

For each new job:

1. Hard gates: remote, visa, seniority, dealbreakers → fail fast
2. Semantic score: profile sections vs job sections
3. Adjudication: LLM check on borderline fits

Only strong matches should consume tailoring compute and your review time. Read semantic matching vs keyword stuffing for why.

Your job: Skim match insights — skip roles that feel wrong even if scored high.

Phase 3: Tailor (automated draft, human edit)

Per match, generate:

• Resume PDF — emphasis aligned to JD, same facts
• Cover letter PDF — 2–3 cited evidence links to company needs
• Form answers — mapped to captured field labels

Constraints: truth pledge — no invented employers, dates, or metrics.

Your job: 5–10 minute review per package. Fix tone, cut a bullet, verify work auth answers.

Phase 4: Apply (manual submit)

PrismApply does not auto-submit. You:

1. Open apply URL from dashboard
2. Upload PDFs or paste fields
3. Mark application as sent for tracking

Why manual submit? Accountability, accuracy, and avoiding auto-applier reputation damage.

Target throughput: 3–8 quality applications per week alongside interview prep — not 50 sloppy ones.

Phase 5: Interview prep (where time goes)

Reclaimed apply time should move to prep:

• Company research
• System design refresh
• Story rehearsal using same profile stories
• Mock interviews

The product promise: you focus on interviews; we keep searching and tailoring in the background.

Weekly rhythm

Day	Activity
Mon	Review new matches in dashboard
Tue–Thu	Apply to 1–2 packages per day
Fri	Profile updates (new project? skill ramp?)
Ongoing	Discovery + matching run automatically

Metrics to track

• Packages reviewed / week
• Applications sent / week
• Reply rate
• Phone screens / month

If reply rate is low despite tailored packages, tighten match thresholds or improve profile evidence — not application volume.

Common mistakes

1. Skipping profile depth → thin tailoring
2. Applying to every match → burnout
3. Using auto-appliers for volume → reputation risk
4. Letting AI invent bullets → interview failure
5. No tracking → lose follow-up threads

Tools checklist

• Profile + tailoring: PrismApply (free beta)
• ATS targets: Lever, Greenhouse, Ashby
• Optional: spreadsheet or dashboard for recruiter threads

Role-specific landing

Deep dive for SWEs: /for/software-engineers

Compare approaches: /vs/auto-apply

Start the playbook

Build your profile once — free

You will not automate your way to an offer — but you can automate everything around the apply button so your energy goes where it counts: proving you can do the job.