Exo ⓧ

We Build Across AI, Blockchain & Product Engineering

[Exo ⓧ @exo_tech_]

Are you a talented developer? Do you want to build products that have a massive impact?

If you're mission driven, high agency, ship outcomes, and don't shy away from challenging problems, then you could be the next engineer on our team.

Apply here: https://t.co/tcg79naKT8

[Exo ⓧ @exo_tech_]

Teach our team OR pitch something if it's really subtle and interesting :)

[Taylor2022 ⓧ @NFTtaylor]

We've been doing lunch and learns @exo_tech_ where anyone from the team can present on a topic they're interested in.

Anyone external interested in presenting to our team? Could be any topic you find interesting.

Some topics we've covered:
- Hazelnut farming
- Bordeaux wine (investor perspective)
- Machine Learning
And much more

[Exo ⓧ @exo_tech_]

DeFi Dev Corp holds over 2 million SOL.

Not a fund. A publicly traded company on Nasdaq (DFDV) built entirely around accumulating and compounding Solana.

We talked to their COO @TheOtherParker_ about what that actually looks like from the inside.

@DefiDevCorp

[Exo ⓧ @exo_tech_]

Traditional finance is hiring blockchain engineers faster than crypto companies right now.

Fidelity, BlackRock, JPMorgan, Baillie Gifford. All building on-chain infrastructure.

The skillset they need: engineers who understand both TradFi compliance and DeFi architecture. That intersection is tiny. Demand is massive.

[Exo ⓧ @exo_tech_]

8/ The compounding effect: a 30% reduction in CU usage across a protocol doing 100K transactions per day saves real money in priority fees. At scale, optimization is not a nice-to-have. It is a cost structure decision.

[Exo ⓧ @exo_tech_]

1/ Solana transaction optimization is not about writing clever code. It is about understanding how the runtime executes your instructions and eliminating the work it does not need to do. Here is a technical breakdown.

[Exo ⓧ @exo_tech_]

7/ Preflight simulation: always simulate before submitting. It catches errors before they cost you fees and lets you fine-tune your compute budget based on actual usage.

[Exo ⓧ @exo_tech_]

The embedded team model works because context is the most expensive thing in engineering.

Our engineers join your Slack, your standups, your repo. They do not hand off code. They are in the architecture decisions.

Clients stay because the model works. Not because of a contract. That is trust.

[Exo ⓧ @exo_tech_]

sRFC 40: ERC-4626 for Solana.

One standard interface for tokenized vaults. Deposits, withdrawals, share accounting.

We submitted it to the Solana Foundation because building custom vault integrations for every protocol is not sustainable. Standards compound. Custom code does not.

[Exo ⓧ @exo_tech_]

We're sunsetting Exo's AI auditing tool.

We built it, shipped it, and learned a ton from the teams who tried it. The biggest takeaway: in smart contract security, the stamp of approval matters more than the tooling. That requires a full auditing practice, not just software.

The AI and agentic tech behind it isn't going away. It's now part of how our engineering team works internally, helping us review code faster and more thoroughly for the teams we build with.

More products coming. This was the first, not the last.

Thanks to everyone who tested it and gave feedback.

[Exo ⓧ @exo_tech_]

7/ The teams that get these patterns right early save months of rework later. MCP is infrastructure. Treat it with the same rigor you give your REST APIs and database schemas.

Reach out to talk: https://t.co/speqmGE6KX

[Exo ⓧ @exo_tech_]

1/ MCP servers are becoming critical infrastructure for enterprise software. But most teams are building them wrong. Here are 5 patterns we have seen work across production deployments.

[Exo ⓧ @exo_tech_]

6/ Pattern 5: Versioned Schemas. Your MCP server will evolve. Version your tool schemas from day one. Breaking changes to tool interfaces break every AI integration that depends on them. Treat your MCP schema like a public API contract.