Working with me — a strategy.

Generalist full-stack engineers are a commodity. The wedge is being the person who builds multi-tenant SaaS, async fulfillment, and three-tier products — and has the shipped projects to prove it.

What I lead with (in this order)

1. Multi-tenant by default — Ticmint, Atlas, Energy TMS all have a tenant layer. Most freelancers can't honestly say that.
2. Async + queue-backed fulfillment — BullMQ + Kafka shipped to production. Specific, not theoretical.
3. Three-tier products — admin web + API + technician mobile app, all in one codebase. Rare combination.
4. Regulatory + payment flows — DTCM, Stripe webhooks, KYC-style flows actually in the wild.

Who I'm not for

• Pure design jobs or visual-only work.
• One-week "build me a landing page" briefs.
• Heavy data science / ML model work (I integrate, I don't train).
• Anything where "AI agent" is the entire product and not the feature.

Being clear about the no sharpens the yes. It also saves both sides a discovery call.

Three concrete archetypes. Most inbound either matches one of these or politely shouldn't.

For each ICP, two things to know cold: the title of the buyer (usually founder/CEO for A, CTO for B, COO for C) and their existing pain (Series A pressure for A, "we're rebuilding for the third time" for B, "tickets in WhatsApp" for C).

Three productised offers and a discovery sprint. Fixed prices on the small tier, day-rate ranges on the large ones. Never quote without scoping; never estimate without the discovery sprint.

Payment terms

• 50% on signature, 50% on delivery. Or thirds for engagements > 8 weeks.
• Net-7, late fee 2% / week after.
• USD via Wise / Payoneer (international). Bank transfer fine for India clients.
• Always sign a one-page MSA + per-project SOW. Always.

Pick two, do them weekly for a year, ignore the rest. Most independents fail because they sprinkle effort across eight channels instead of compounding on two.

Three content formats, three publish slots a week, one master rule: everything ladders back to a shipped project. No abstract takes.

Weekly cadence

• Monday morning: short Format A post (LinkedIn + X).
• Wednesday morning: Format C carousel or thread (LinkedIn).
• Friday afternoon: Format A post or repost-with-comment of an ICP-fit person's content.
• First Tuesday of month: Format B deep-dive published on site + announced on LinkedIn.

Topic backlog (drawn from real work)

• Multi-tenant data models in NestJS — three patterns and when each breaks.
• Why BullMQ wins over Kafka for most SaaS fulfillment (and when it doesn't).
• Integrating DTCM regulatory ticketing without blocking checkout.
• Token-based public invoices: a 50-line auth bypass that's actually safe.
• Angular 17 + NestJS 11 — a starter monorepo I keep forking.
• Three-tier products: how to scope admin + API + mobile in one sprint.
• Ionic + Capacitor 8 plugins that ship 90% of "we need a mobile app" briefs.
• Where OpenAI in a CRM actually earns its keep — and where it's theatre.

Outbound to 15 ICP-fit founders / CTOs per week. Hyper-targeted, never templated-feeling. Goal isn't a sale — it's a single "tell me more" reply.

Targeting

• LinkedIn Sales Navigator filters: Title (Founder / CTO / Head of Engineering) + Company size (2–30) + Industry (B2B SaaS, agency, field service) + Hiring "engineer" recently.
• Trigger events to look for: just raised seed/A · just posted "hiring full-stack" · just announced a rebuild · their CTO just left.
• Disqualifiers: 50+ engineers (too big), pre-revenue idea-stage (wrong fit), heavy ML / data-science focus.

First-touch DM (LinkedIn)

Why this works: specific trigger, specific proof, specific technical observation, no ask. They reply because you're showing competence, not selling. Reply rate target: 18–25%.

Cold email (when you have an intro reason)

Follow-up cadence

1. Day 0: first touch (DM or email).
2. Day 5: one follow-up if no reply — a piece of new value, not a "just checking in". E.g. link to the case study you mentioned.
3. Day 21: final follow-up — share a relevant post you wrote since.
4. Day 90: add them to a quarterly "what I shipped" newsletter (with explicit opt-in).

Three touches max. Then leave them alone. Annoying outbound is worse than no outbound.

Discovery → call → proposal → close. Each stage has one job. Don't skip; don't merge.

1. Discovery (email exchange, 2–4 messages): What are they building? What's the actual scope? Budget range? Timeline? Anyone else evaluated? If any answer is missing or fuzzy, do not jump to a call.
2. Intro call (30 min, free): Confirm fit. Ask three things: what does success look like in 90 days · what's the cost of not doing this · who else is involved in the decision. End with: "Here's what I think the right shape is. I'll send a discovery-sprint proposal by tomorrow if you're in."
3. Discovery sprint proposal (sent within 24h): One page. Scope, deliverables, price, timeline, payment terms. Decision date 5 days out.
4. Discovery sprint (1 week, paid): Architecture doc, milestones, risks, fixed-price quote for the main engagement. Delivered as a real document, not a Notion page.
5. Main engagement proposal: Built on the discovery sprint output. Pre-sold by then because they trust the doc.
6. Sign & kickoff: Contract within 48h of yes. Kick off within a week of signature.

Concrete weekly actions to go from "I have a portfolio" to "I have inbound deals and a content pipeline that runs without me." 12 weeks.

The minimum set of documents and assets you need before any of this works. Build these once, reuse them on every deal.

• Portfolio site — this one. Custom domain, case studies, services, contact. ✓ Done.
• One-page case study PDFs — one per project. Used as email attachments, dropped in LinkedIn DMs, sent before intro calls.
• Discovery sprint proposal template — one page. Scope, deliverables, price, timeline, payment terms. Reusable for every deal.
• Main engagement proposal template — 3–4 pages. Built from discovery sprint output. Includes scope, milestones, payment schedule, change-order policy.
• MSA + SOW templates — one master service agreement + per-project statement of work. Have a lawyer review once. Reuse forever.
• Intro call agenda (Notion / Loom) — 30-min structure. Stops calls from drifting.
• "How I work" page — this document. Send the link before every intro call. Massive trust signal.
• Stack one-pagers — short PDF for each stack (NestJS + Angular SaaS, Ionic mobile, three-tier field-ops). Used as inbound magnets.
• Quarterly newsletter — opt-in only, 4×/year. "Here's what I shipped, here's what I'm taking on next." Becomes referral fuel.
• Resume PDF — for the rare ICP that asks. Don't lead with it.

What to track weekly. Hide everything else. Five numbers tell you whether the system is working.

The goal of the first year isn't to maximise revenue — it's to make the inbound % climb. By month 18 most deals should come to you, not from outbound. That's when the system actually works.

Three markets, three sub-playbooks. Most India-based independents fail to land international clients because they treat "Western buyer" as one market — it isn't. Different timezones, different buying culture, different invoice + tax friction. Here is the per-region adjustment.

United Kingdom 🇬🇧 — best first market

Netherlands 🇳🇱 — second market

United States 🇺🇸 — biggest pond, hardest first cast

Weekly outbound calendar (multi-region)

Once you've cracked one region, layering a second is timezone-driven:

Payment + invoicing stack

• Wise Business — multi-currency receive (USD, GBP, EUR), local-bank-details in each. Required, not optional.
• Payoneer — backup for clients who can't use Wise. Slightly worse rates.
• Stripe Invoicing — for one-off small US clients who want to pay by card. Fees high but conversion is too.
• W-8BEN form — keep a signed PDF ready to send US clients before any invoice.
• Single invoice template — your name, GST/PAN, their company + VAT number, line items, due date (Net-7), Wise + Payoneer + bank details in footer.
• Currency rule — always invoice in the client's local currency, never in INR. Eats the FX risk yourself; it's worth the trust signal.

Every client in 2026 is asking how you use AI. The right answer isn't "I do everything by hand." It's "I'm the person reviewing the AI." Here's how to position, deliver, and price that.

My actual workflow

What it actually speeds up

• Discovery sprints — architecture docs written in 1 day instead of 3. Tradeoffs explored more thoroughly because exploration is cheap.
• Boilerplate-heavy modules — CRUD endpoints, form scaffolds, type definitions. The mechanical parts of NestJS + Angular.
• Test coverage — generating the boring 80% of tests so you can focus on the edge cases.
• Documentation — README, API docs, runbooks, ADRs. Always slow without AI, fast with it.
• Code review feedback — Claude reviews your own PRs before you push them. Catches embarrassments.
• Migrations + refactors — multi-file renames, schema migrations, framework upgrades. The "tedious but tractable" work.

What it explicitly does NOT change

• The architectural decisions — those stay with me. AI helps explore, doesn't decide.
• Production correctness — every line is reviewed before it lands. No "AI wrote this, ship it."
• Security-sensitive code — auth, payments, regulatory flows. Manually written or carefully reviewed.
• The discovery conversation — clients hire the engineer who understood their business, not the AI that drafted code.

Pricing implication

Same pricing, faster timelines. Don't discount because you're using AI — you're delivering the same shipped thing in less calendar time. That's a benefit to the client (faster time-to-revenue), not a cost reduction for them. If anything, you can charge a small premium for tighter timelines.

If a client pushes back: "the AI tools cost money, the workflow took years to build, and the value you're getting is the shipped feature on time — not the hours behind it."

How to bring it up on sales calls

1. Pre-empt the question — bring up AI usage in the intro call yourself, before they ask. Shows you're not hiding it.
2. Be specific — name the tools (Claude, Cursor, GPT-5). Vagueness sounds defensive.
3. Set the boundary — "AI handles the mechanical 30%. The architecture, the tradeoffs, the production correctness — that's me."
4. Show, briefly — if relevant, show a 30-second screen-share of you reviewing a Claude-generated PR. Demystifies the workflow instantly.