# Cover Letter Library — Alexander Vasilev ("Lance")

> **What this is:** not one cover letter — a library of interchangeable blocks plus five fully
> assembled, ready-to-send letters for the four engagement types on the resume (Fractional CTO,
> Technical co-founder, Staff/Principal Architect, Investment & strategic partnership) plus a
> short-form reply for recruiter/LinkedIn inbound. Pick the blocks that fit, fill in the
> brackets, cut anything you don't need. **Never send a block combo unedited — see §6.**
>
> **English-only, matching the résumé's own decision** (`CV.htm`, Phase M): this library targets
> international recruiting/consulting/investor audiences. If a specific, high-value opportunity
> is explicitly Russian- or Portuguese-speaking, translate the *assembled* letter for that one
> case rather than maintaining parallel block libraries — same anti-duplication logic as the
> résumé's 3-copy problem.
>
> **Maintainer:** `npm run cv:sync -w agentstack-frontend` · `npm run audit:cv` (CV Plane — `docs.founder.cv_plane.gen1`)
>
> **Kept in sync with:** `CV.htm` v<!-- cv:resume_version -->1.2.0<!-- /cv:resume_version --> (<!-- cv:verified_date -->2026-09-10<!-- /cv:verified_date -->) — numbers, contact details, and location
> here (Santos, São Paulo) must match the résumé. If you update a fact in the résumé, grep this
> file for the old value before you consider the résumé edit "done."

---

## 0. Quick decision guide

| Situation | Hook | Positioning | Proof | CTA | Target length | Channel |
|---|---|---|---|---|---|---|
| Applying to a posted Staff/Principal Architect role | A1 | B1 (+ B4 if career-story-friendly JD) | C1 | D1 | 250–320 words | Email / ATS form |
| Warm intro via mutual contact | A2 | B2 | C2 | D1 or D5 | 180–250 words | Email |
| Cold outreach — company has no posted role but clearly needs senior technical leadership | A3 | B5 | C1 or C2 | D2 | 200–280 words | Email / LinkedIn InMail |
| Technical co-founder conversation | A2 or A3 | B6 | C2 | D3 | 220–300 words | Email |
| Investor / strategic partner (AgentStack itself) | A4 | B3 (if Web3-focused) or B2 | C2 or C3 | D4 | 220–320 words | Email |
| Recruiter reached out first (LinkedIn/email) | A5 | one line only, no full block | C2 (one line) | D5 | 60–110 words | LinkedIn / short email |
| Crypto/Web3-specific role or grant reviewer | A1/A3 | B3 | C3 | D1/D2 | 220–300 words | Email |
| Staff/Principal + reliability / incident-grade systems | A1 | B7 (+ B1) | C1 | D1 | 250–320 words | Email |

Rule of thumb: **one hook, one positioning block, one proof block, one CTA.** Two positioning
blocks is the maximum before a letter reads as unfocused — this is the same "Clarity" failure
mode the résumé itself was audited for.

---

## 1. Subject line bank

Email subject lines convert better when they're specific, not clever. Use `[ ]` fields.

**Job application:**
- `Application: [Role] — Alexander Vasilev (15+ yrs, solo-shipped AI platform)`
- `[Role] at [Company] — AI systems architect, ex-game industry`

**Warm referral:**
- `[Mutual Contact] suggested I reach out — [Role/Topic]`

**Cold outreach (no posted role):**
- `A architecture question about [Company]'s [specific product/system]`
- `Fractional CTO availability — AI/backend systems`

**Co-founder:**
- `Technical co-founder conversation — AI systems background`

**Investor / partnership:**
- `AgentNet (AgentStack) — testnet evidence + partnership conversation`

**Recruiter reply:**
- `Re: [Role] — quick background + availability`

---

## 2. Modular blocks

Each block is written to stand alone as one paragraph. Swap the bracketed tokens; don't stack
more blocks from the same letter than the decision table recommends.

### A. Opening hooks

**A1 — Responding to a specific posted role**
> I'm writing about the **[Role]** opening at **[Company]** [— found via `[source, e.g. your
> careers page / a referral / LinkedIn]`]. [One sentence tying a specific thing about their
> product/stack to your background, e.g.: "Building a system that has to keep AI in the loop of
> real production decisions is exactly the kind of problem I've spent the last two years solving
> at my own company."]

**A2 — Warm referral / mutual contact**
> **[Mutual Contact]** mentioned you're [hiring for `[role]` / looking for technical leadership
> for `[project]`], and suggested I reach out directly. I run **AgentStack.tech**, a
> multi-tenant AI platform I designed and shipped solo, and I'd like to explore whether my
> background is a fit for what you're building.

**A3 — Cold outreach, no posted role**
> I've been following **[Company]**'s work on `[specific product/system/problem]`, and it's the
> kind of system I've spent 15+ years building — [distributed backends / real-time systems / AI
> infrastructure that has to behave predictably under real load]. I'm not responding to a
> specific posting; I'm reaching out because I think there's a fit worth a short conversation.

**A4 — Investor / strategic partner note**
> I'm the founder of **AgentStack.tech**, an AI-first multi-tenant platform with an on-chain
> economy layer (**AgentNet**) currently at the testnet-evidence stage across three chains. I'm
> reaching out because `[fund/person]`'s focus on `[AI infra / Web3 infra / dev tools]` overlaps
> directly with what we've built, and I'd like to share the evidence pack and discuss fit.

**A5 — Recruiter inbound reply**
> Thanks for reaching out about **[Role]** — appreciate you thinking of me. Quick background:
> [one-line summary — see Positioning block, one sentence only].

---

### B. Positioning / value paragraphs (pick one — two only if they don't overlap)

**B1 — AI systems architect / production AI**
> My focus is making AI carry *real* production load — not a chatbot bolted onto CRUD, but
> systems where agent actions are scoped, observable, and reversible by design. At AgentStack I
> built a unified MCP + REST catalog of **<!-- cv:mcp_actions_short -->550+<!-- /cv:mcp_actions_short --> actions** (<!-- cv:mcp_domains -->48<!-- /cv:mcp_domains --> domains, single `agentstack.execute` entry) so AI agents and human developers
> operate the exact same platform, with Field Access Policy, RBAC, and structured instruction
> design (prevention-over-cure prompting, not prompt patching after incidents) governing what AI
> is allowed to touch.

**B2 — Solo-founder / AI-augmented workflow differentiator**
> I designed and built AgentStack **solo** — from an empty repository to a multi-tenant platform
> with <!-- cv:mcp_actions_short -->550+<!-- /cv:mcp_actions_short --> catalog actions and <!-- cv:adr_count -->375<!-- /cv:adr_count --> documented architecture decisions — without a dedicated
> engineering team, using an AI-agent-augmented workflow (structured prompts, codebase
> navigation conventions, controlled-change discipline) as a genuine force multiplier rather than
> a shortcut. For a role built around AI systems, I think that process is itself relevant
> evidence, not just the resulting product.

**B3 — Web3 / AgentNet economy specialist**
> Beyond the AI stack, I designed and shipped **AgentNet**, an on-chain economy layer with an
> append-only ledger as financial source of truth and Solidity contracts (ERC-4626 vault,
> ERC-8004 proof registries) as public anchors — not a token bolted onto a database. It's
> currently at the testnet/pre-mainnet evidence stage across Base Sepolia, BSC testnet, and
> Arbitrum Sepolia, with a source-verified contract and a public multi-chain proof timeline.

**B4 — Career arc: games → distributed backend → AI infra**
> My background starts in game development — multiplayer netcode, performance-sensitive systems,
> shipping on fixed console-certification deadlines at studios in Stockholm and Oslo — which is
> where I picked up the discipline of building for correctness under real constraints, not just
> for demos. That discipline carried directly into distributed backends and, most recently, into
> AI infrastructure that has to stay predictable as usage grows.

**B5 — Fractional CTO specific value**
> I work best embedded but time-boxed: fast architectural diagnosis in the first few weeks,
> a concrete remediation or build plan, then either hands-on delivery or a clean handoff to your
> team — not an open-ended retainer. I've done this for my own platform at a scale (<!-- cv:mcp_actions_short -->550+<!-- /cv:mcp_actions_short --> API
> surface, multi-tenant data model) that most fractional engagements never reach solo.

**B7 — Organism reliability / platform ops (Staff/Principal, Fractional CTO)**
> Beyond feature shipping, I design for incident-grade operability: unified messenger history (no split-brain chat paths), 8DNA leaf write discipline (no fat-blob data wipes), and a Dev Test Atlas (`test:scope`) so subsystems stay diagnosable under load — the same reliability plane we shipped in AgentStack 0.4.17–0.4.18, not demo-only architecture.

**B6 — Technical co-founder specific value**
> I'm looking for a founding team where I own architecture and delivery end-to-end while a
> co-founder owns `[go-to-market / domain expertise / capital]` — the same split that let me
> ship AgentStack's entire technical surface without a hired engineering team. I bring shipped,
> load-bearing systems (not a prototype) as proof I can execute alone when needed, which is
> exactly the risk a co-founder relationship is meant to de-risk.

---

### C. Proof paragraphs (pick one)

**C1 — Engineering-discipline flavor (for technical hiring managers / CTOs)**
> Concretely: <!-- cv:adr_count -->375<!-- /cv:adr_count --> Architecture Decision Records (one documented decision per major subsystem,
> not tribal knowledge), <!-- cv:audit_scripts -->123<!-- /cv:audit_scripts --> root CI audit scripts enforcing architectural invariants beyond unit tests, and
> <!-- cv:e2e_specs_total -->118<!-- /cv:e2e_specs_total --> Playwright specs (<!-- cv:e2e_specs_frontend -->0<!-- /cv:e2e_specs_frontend --> in frontend/e2e — methodology in cv.json). I'd rather show
> you the paper trail than ask you to take "production-grade" on faith.

**C2 — Scale & shipped-surface flavor (for founders / business audience)**
> Concretely: a live platform with <!-- cv:mcp_actions_short -->550+<!-- /cv:mcp_actions_short --> catalog actions spanning auth, billing, real-time
> social, commerce, and an on-chain economy layer — built and operated solo. You can see it
> running, not just read about it: `[link a specific live proof relevant to them, e.g.
> agentstack.tech/grants/demo or agentstack.tech/s/2/demo-store/]`.

**C3 — Web3-specific proof**
> Concretely: a source-verified AGNT contract on Base Sepolia
> (`sepolia.basescan.org/address/0x38aD…9a943`, compiler v0.8.24) and a public multi-chain proof
> timeline at `agentstack.tech/grants/demo` — testnet evidence you can check yourself, not a
> whitepaper claim.

---

### D. Closing / CTA variants

**D1 — Request interview/call (job application)**
> I'd welcome the chance to talk through how this maps to `[Role]` — happy to work around your
> schedule. My full background, with links to live proof and a machine-readable summary, is at
> **agentstack.tech/CV.htm** (`cv.json` for the structured version).

**D2 — Propose short diagnostic call (fractional CTO)**
> If it's useful, I'd suggest a 30-minute call — no obligation — where I ask enough questions to
> tell you honestly whether a fractional engagement makes sense for where you are right now.
> Full background: **agentstack.tech/CV.htm**.

**D3 — Open a co-founder conversation**
> If this resonates, I'd like to start with a conversation, not a commitment — what you're
> building, what you're looking for in a technical co-founder, and whether there's a real fit
> before either of us talks about terms. Background: **agentstack.tech/CV.htm**.

**D4 — Request investor call / deck exchange**
> I'd like to share the full evidence pack (grant audit pack, live demo, contract addresses) and
> find 20–30 minutes to talk through where AgentStack fits your thesis. Happy to send materials
> ahead of a call.

**D5 — Short, low-pressure sign-off (cold outreach / LinkedIn)**
> No pressure either way — happy to chat if it's a fit, and happy to be told it isn't. Background
> here if useful: **agentstack.tech/CV.htm**.

---

## 3. Fully assembled letters

Placeholders in `[brackets]` — fill in before sending. Word counts are for the assembled body
text only (excluding greeting/signature).

### 3.1 — Fractional CTO outreach (A3 + B5 + C1 + D2) · ~230 words

> Subject: Fractional CTO availability — AI/backend systems
>
> Hi `[Name]`,
>
> I've been following `[Company]`'s work on `[specific product/system/problem]`, and it's the
> kind of system I've spent 15+ years building — distributed backends and, more recently, AI
> infrastructure that has to behave predictably under real load. I'm not responding to a specific
> posting; I'm reaching out because I think there's a fit worth a short conversation.
>
> I work best embedded but time-boxed: fast architectural diagnosis in the first few weeks, a
> concrete remediation or build plan, then either hands-on delivery or a clean handoff to your
> team — not an open-ended retainer. I've done this for my own platform, AgentStack, at a scale
> (<!-- cv:mcp_actions_short -->550+<!-- /cv:mcp_actions_short --> API surface, multi-tenant data model) that most fractional engagements never reach solo.
>
> Concretely: <!-- cv:adr_count -->375<!-- /cv:adr_count --> Architecture Decision Records, <!-- cv:audit_scripts -->123<!-- /cv:audit_scripts --> CI scripts enforcing architectural
> invariants beyond unit tests, and <!-- cv:e2e_specs_total -->118<!-- /cv:e2e_specs_total --> end-to-end test specs across the platform I built and
> still operate. I'd rather show you the paper trail than ask you to take "production-grade" on
> faith.
>
> If it's useful, I'd suggest a 30-minute call — no obligation — where I ask enough questions to
> tell you honestly whether a fractional engagement makes sense for where you are right now. Full
> background: **agentstack.tech/CV.htm**.
>
> Best,
> Alexander Vasilev
> `[signature block — see §7]`

---

### 3.2 — Technical co-founder (A2 + B6 + C2 + D3) · ~240 words

> Subject: Technical co-founder conversation — AI systems background
>
> Hi `[Name]`,
>
> `[Mutual Contact]` mentioned you're looking for a technical co-founder for `[project/idea]`,
> and suggested I reach out directly.
>
> I'm looking for a founding team where I own architecture and delivery end-to-end while a
> co-founder owns `[go-to-market / domain expertise / capital]` — the same split that let me ship
> AgentStack's entire technical surface without a hired engineering team. I bring shipped,
> load-bearing systems as proof I can execute alone when needed, which is exactly the risk a
> co-founder relationship is meant to de-risk.
>
> Concretely: a live platform with <!-- cv:mcp_actions_short -->550+<!-- /cv:mcp_actions_short --> catalog actions spanning auth, billing, real-time social,
> commerce, and an on-chain economy layer, built and operated solo, with <!-- cv:adr_count -->375<!-- /cv:adr_count --> documented
> architecture decisions behind it. You can see it running, not just read about it:
> **agentstack.tech**.
>
> If this resonates, I'd like to start with a conversation, not a commitment — what you're
> building, what you're looking for in a technical co-founder, and whether there's a real fit
> before either of us talks about terms. Background: **agentstack.tech/CV.htm**.
>
> Best,
> Alexander Vasilev
> `[signature block — see §7]`

---

### 3.3 — Staff/Principal Architect job application (A1 + B1 + B4 + C1 + D1) · ~300 words

> Subject: Application: `[Role]` — Alexander Vasilev (15+ yrs, solo-shipped AI platform)
>
> Dear `[Hiring Manager Name / Hiring Team]`,
>
> I'm writing about the `[Role]` opening at `[Company]` [— found via `[source]`].
> `[One sentence tying something specific about their product/stack to your background.]`
>
> My focus is making AI carry real production load — not a chatbot bolted onto CRUD, but systems
> where agent actions are scoped, observable, and reversible by design. At AgentStack, the
> platform I founded and architected, I built a unified MCP + REST catalog of <!-- cv:mcp_actions_short -->550+<!-- /cv:mcp_actions_short --> actions so AI
> agents and human developers operate the exact same platform, governed by Field Access Policy,
> RBAC, and structured instruction design.
>
> My background starts in game development — multiplayer netcode and performance-sensitive
> systems shipped on fixed console-certification deadlines — which is where I picked up the
> discipline of building for correctness under real constraints. That discipline carried directly
> into distributed backends and, most recently, into AI infrastructure that has to stay
> predictable as usage grows.
>
> Concretely: <!-- cv:adr_count -->375<!-- /cv:adr_count --> Architecture Decision Records, <!-- cv:audit_scripts -->123<!-- /cv:audit_scripts --> CI scripts enforcing architectural
> invariants beyond unit tests, and <!-- cv:e2e_specs_total -->118<!-- /cv:e2e_specs_total --> end-to-end test specs. I designed and shipped this solo,
> using an AI-agent-augmented workflow as a force multiplier — itself relevant experience for a
> role built around AI systems.
>
> I'd welcome the chance to talk through how this maps to `[Role]` — happy to work around your
> schedule. My full background, with links to live proof and a machine-readable summary, is at
> **agentstack.tech/CV.htm** (`cv.json` for the structured version).
>
> Best regards,
> Alexander Vasilev
> `[signature block — see §7]`

---

### 3.4 — Investor & strategic partnership note (A4 + B3 + C3 + D4) · ~250 words

> Subject: AgentNet (AgentStack) — testnet evidence + partnership conversation
>
> Hi `[Name]`,
>
> I'm the founder of AgentStack.tech, an AI-first multi-tenant platform with an on-chain economy
> layer — AgentNet — currently at the testnet-evidence stage. I'm reaching out because
> `[fund/person]`'s focus on `[AI infra / Web3 infra / dev tools]` overlaps directly with what
> we've built, and I'd like to share the evidence pack and discuss fit.
>
> AgentNet has an append-only ledger as financial source of truth and Solidity contracts (ERC-4626
> vault, ERC-8004 proof registries) as public anchors — not a token bolted onto a database. It
> runs across three testnets (Base Sepolia, BSC testnet, Arbitrum Sepolia) via a single chain
> orchestration layer, and it sits on top of a platform with <!-- cv:mcp_actions_short -->550+<!-- /cv:mcp_actions_short --> shipped catalog actions and <!-- cv:adr_count -->375<!-- /cv:adr_count -->
> documented architecture decisions — the whole thing designed and built solo, with an
> AI-agent-augmented workflow as the operating method rather than a marketing line.
>
> Concretely: a source-verified AGNT contract on Base Sepolia
> (`sepolia.basescan.org/address/0x38aD…9a943`, compiler v0.8.24) and a public multi-chain proof
> timeline at `agentstack.tech/grants/demo` — evidence you can check yourself.
>
> I'd like to share the full evidence pack and find 20–30 minutes to talk through where AgentNet
> fits your thesis. Happy to send materials ahead of a call.
>
> Best,
> Alexander Vasilev
> `[signature block — see §7]`

---

### 3.5 — Short LinkedIn / recruiter reply (A5 + one-line B + one-line C + D5) · ~90 words

> Thanks for reaching out about `[Role]` — appreciate you thinking of me.
>
> Quick background: I'm the founder of AgentStack, an AI platform I designed and built solo —
> <!-- cv:mcp_actions_short -->550+<!-- /cv:mcp_actions_short --> API actions, <!-- cv:adr_count -->375<!-- /cv:adr_count --> documented architecture decisions, plus an on-chain economy layer
> (AgentNet) at the testnet stage. Before that, 15+ years in distributed backends and game
> development (Stockholm, Oslo).
>
> No pressure either way — happy to chat if it's a fit. Full background:
> **agentstack.tech/CV.htm**.
>
> Best,
> Alexander

---

## 4. Best-practice checklist (industry standards)

Apply this to *every* assembled letter before sending, not just the templates above:

- [ ] **Under ~300 words / 4 short paragraphs.** Longer than that and it stops getting read in full.
- [ ] **Opens with a hook, not "I am writing to apply for..."** — the reader already knows why the email exists; spend the first sentence on relevance instead.
- [ ] **One clear differentiator, not a list.** Pick the *one* thing (solo-built platform, AI-augmented workflow, Web3 evidence, game-industry discipline) that matters most for *this* reader — don't try to fit all of them in.
- [ ] **Quantified proof, not adjectives.** "<!-- cv:mcp_actions_short -->550+<!-- /cv:mcp_actions_short --> actions, <!-- cv:adr_count -->375<!-- /cv:adr_count --> ADRs, <!-- cv:e2e_specs_total -->118<!-- /cv:e2e_specs_total --> e2e specs" beats "extensive experience" every time — this mirrors the résumé's own "number without a source is not a fact" rule.
- [ ] **Mirror the job posting's actual language** where honest — if they say "distributed systems," don't substitute a synonym; ATS keyword matching and human skimming both reward literal overlap.
- [ ] **Always link the résumé and, where relevant, one live proof URL** — never assume they'll go looking.
- [ ] **Plain-text safe.** Copy the assembled letter into a plain email body or a plain-text field, not a styled document, unless a specific application system asks for an attachment. No tables, no exotic Unicode bullets, no emoji.
- [ ] **No clichés:** avoid "team player," "passionate," "synergy," "results-driven," "go-getter." Every sentence should survive the question "could literally anyone claim this?" — if yes, cut or make it specific.
- [ ] **Proofread every bracket.** A leftover `[Company]` or wrong pronoun is an instant signal of a mass-sent template — which this genuinely is, at the block level, so the fill-in pass matters more, not less.
- [ ] **Match tone to channel:** email/application-form letters can run the full ~250–300 words; LinkedIn messages and recruiter replies should stay under ~110 words — nobody reads a full cover letter in a chat UI.
- [ ] **Follow up once, 5–7 business days later,** with a one-line bump ("Following up on this — still interested if there's a fit"), not a re-send of the same letter.

---

## 5. Web3-safety note (carried over from the résumé's own rule)

If a letter touches AgentNet/Web3 (blocks B3, C3, or the investor letter), keep the same
discipline the résumé enforces (`CV.htm` Phase L):

- Always state the **testnet / pre-mainnet** status explicitly — never imply a live financial
  product.
- Never use "invest," "returns," or "yield for users" in a marketing sense — use engineering
  framing ("vault accounting mechanics," "reference implementation," "testnet economy design").
- Use only current terminology — **AgentNet**, **AGNT**, **agUSD** — never the retired
  "AgentChain" label.
- Re-verify the testnet contract address/status before sending anything that cites it; testnet
  contracts can be redeployed.

---

## 6. Customization checklist — never send a block combo unedited

Before sending, confirm you've replaced or removed every placeholder:

- [ ] `[Company]`, `[Role]`, `[Hiring Manager Name]` filled in correctly (spelled, capitalized).
- [ ] `[Mutual Contact]` named only if you have their actual permission to name-drop.
- [ ] `[source]` (how you found the role) is true, not generic.
- [ ] The one "tie-in sentence" in A1/A3/A4 references something *specific* to that company/fund
      — a product, a blog post, a talk, a stated thesis. If you can't write one honest specific
      sentence, that's a signal to research more before sending, not to skip the sentence.
- [ ] The live-proof link you chose (grants/demo, hosted storefront, etc.) is actually relevant
      to the reader's domain — don't send a Web3 proof link to a non-crypto hiring manager.
- [ ] Numbers match the current `CV.htm` — re-check `<!-- cv:mcp_actions_short -->550+<!-- /cv:mcp_actions_short -->`, `<!-- cv:adr_count -->375<!-- /cv:adr_count -->`, `<!-- cv:audit_scripts -->123<!-- /cv:audit_scripts -->`, `<!-- cv:e2e_specs_total -->118<!-- /cv:e2e_specs_total -->` against the
      résumé's own "verified" comments before reusing them; they are dated, not permanent.

---

## 7. Signature block (reusable)

```
Alexander Vasilev
AI Systems Architect · Founder & Chief Architect, AgentStack.tech
agentstack.tech/CV.htm
lance@world4play.com · +55 (13) 99800-4433 (WhatsApp/phone) · Telegram @LanceW4P
linkedin.com/in/alexanderlance
Santos, São Paulo, Brazil · Remote-first · GMT-3
```

Shorter variant for LinkedIn / chat channels where a full signature is redundant:

```
Alexander Vasilev — agentstack.tech/CV.htm
```

---

<!--
Cover letter library changelog:
- v1.0.0 (2026-07-08): initial modular block library + 5 assembled letters, synced with CV.htm
  v1.1.0 (<!-- cv:mcp_actions_short -->550+<!-- /cv:mcp_actions_short --> actions, <!-- cv:adr_count -->375<!-- /cv:adr_count --> ADRs, 112 audit scripts, <!-- cv:e2e_specs_total -->118<!-- /cv:e2e_specs_total --> e2e specs, Santos location, AgentNet
  testnet status, patent-pending status, solo-architected/AI-augmented positioning).
-->