Evan D'Souza
Solo SaaS founder. Ops brain. Local LLM tinkerer. Music producer. Motorcyclist.
Currently: Building Dszape (hotel SaaS) + Becky OS (multi-agent coding OS)
Previously: 10 years turning early-stage chaos into predictable scale
Always: Figuring it out as I go, honestly

I build things. Sometimes it's a SaaS platform for hotels. Sometimes it's a multi-agent AI system that manages itself. Sometimes it's a Konkani hip-hop track. Sometimes I just need to disappear on a motorcycle for a while.

The common thread: I can't sit still, and I'd rather build something imperfect today than plan something perfect forever.

I also work with early-stage startups as a hands-on growth consultant through evandsouza.com. Not the "here's a deck, good luck" kind -- the "I'll build the engine alongside your team for 90 days" kind.

What I help with: Startup Growth | SaaS Strategy | D2C Operations | Product-Led Growth | Community Building | AI Strategy

"Most startups don't die from bad products. They die from bad plumbing." -- me, probably too often

I don't believe in being one thing. Life's too short and too interesting.

           GREENFIELD MODE                          BROWNFIELD MODE
           ===============                          ===============

     Brief --> PRD --> UX --> Arch             Audit --> Document --> Plan
       |                       |                |                     |
       v                       v                v                     v
     Stories --> Implement --> Test          Intervene --> Test --> Verify
                    |                                       |
                    v                                       v
              [ VERIFIER ]                            [ VERIFIER ]
              DONE | VERIFIED | AUDITED               DONE | VERIFIED | AUDITED
                    |                                       |
                    v                                       v
         Retro --> Rule --> Compile             Retro --> Rule --> Compile
                    |                                       |
                    +---- RULES COMPOUND ------------------+

The agent that builds the feature cannot be the agent that grades it. That's not a technical detail -- it's the whole point.

I'm not going to pretend I have it all figured out. But after years of building, breaking, and rebuilding things, here's what I keep coming back to:

• Systems beat heroics. A good process run by average people outperforms a bad process run by geniuses. Every time.
• Ship the smallest thing that validates the assumption. Then iterate. The 60-second test: can someone get value in under a minute?
• If it's not measured, it didn't happen. If it's not automated, it won't stick.
• The same mistake shouldn't happen twice. Every incident becomes a rule. Rules compound. That's the whole game.
• You can be serious about your work without taking yourself too seriously. I build production SaaS by day and make Konkani hip-hop by night. Both are real.

Things that run on my own machine, because privacy, iteration speed, and ₹0 marginal cost matter:

• Ollama on Apple Silicon — local models for private, repeated, high-iteration workloads
• A persistent AI assistant voice — local TTS / voice-clone experiments, no rented API voice
• Agent long-term memory — file-based hot index + cold archive + full-history recall, injected at session start
• FL Studio MCP server — agents write MIDI straight into my DAW sessions
• Offline AI vocal mastering — autotune + processing chain + loudness, fully local, used on real releases

Local models where privacy and iteration speed matter; frontier models where capability does.

• How I'm Building a Full SaaS Platform Solo With AI at 36
• Running Local LLMs on a MacBook: The Practical Guide for Builders
• What I Learned Building a Multi-Agent AI Coding System (BeckyOS)
• Vibe Coding in Production: Guardrails That Keep AI-Written Code Safe
• Real Reason Why People Contribute to Open Source -- A Business Use Case

Say Hi

I'm always up for conversations about vibecoding, solopreneur SaaS, growth ops, Konkani culture, or motorcycle routes you think I should ride.

"I'd rather build something imperfect today than plan something perfect forever."