


toluhunter@site:~$ whoami
Tolulope Fakoya
The limit of what you can build is near your imagination. The limit of what survives production is infrastructure.
I design, monitor, and secure yours so the platform you ship performs as well as the idea that started it.
I work with small teams shipping an AI or cloud product who don't have a platform team in place yet.
Production-Grade AI Assistant with Built-In Guardrails
Guardrails used for real load, at a price that scales with the your budget.



// the problem
An AI assistant that's actually production-ready has to be cost-effective to run and secure against attacks, including the reworded prompts that slip past a simple filter.
// what i built
A two-layer guardrail graph on LangGraph: a fast regex check catches the obvious attacks before any model call happens, and a semantic LLM check catches the reworded ones, checking both what the user sends and what the model replies with, so a bypass on the way in doesn't help if the reply is still off-policy on the way out.
// the result
Same design, same guardrails, deployed at two different budgets depending on what the client needs.



File Transfer Without Walls
Works identically across every major platform, and no server ever sees your files.
// the problem
AirDrop only works between Apple devices, and Quick Share skips iOS and macOS entirely. Nothing lets you send a file to any device, on any platform, without installing extra software or routing it through someone else's server.
// what i built
A cross-platform transfer app in Rust that works the same way everywhere: install it once and it finds every other device running it, on the same network or over the internet. Online transfers go peer-to-peer over WebRTC, so file data never touches a server, and recipients without the app can still receive files straight through a browser.
// the result
Works the same way on Windows, macOS, Linux, iOS, and Android, with no server ever touching the file data in an online transfer.
Mobile App Infrastructure
// the problem
A large financial institution's mobile app backend couldn't scale or recover on its own, and every spike in load turned into downtime users could feel.
// what i built
Containerized the application and moved it onto Amazon EKS, with pod-level autoscaling so the infrastructure absorbs load spikes and recovers from failures automatically instead of needing a human to step in.
// the result
Downtime over the year dropped by roughly half.
Migration from DigitalOcean to AWS
// the problem
An online marketplace had outgrown DigitalOcean and needed to move onto AWS without losing a single record out of its relational database mid-migration.
// what i built
A full re-platform onto AWS: a load balancer and auto-scaling group in front of the application, RDS and ElastiCache for the data layer, AWS DMS to migrate the relational database live, object storage moved to S3, and CloudFront in front of it all for the CDN.
// the result
The marketplace now runs on infrastructure it didn't have on DigitalOcean, with traffic spikes absorbed by the auto-scaling group and ElastiCache instead of by hand, and CloudFront in front of it for the CDN layer.
Have a project in mind or not sure where to start? Book a free call and we'll talk through what you're building, what problems need solving, and whether I'm the right fit.
Book a Free Callor send me an emailBuy me a coffee