About the service
I'm an Erlang and OTP developer and I build backend systems that keep working when everything around them falls apart. At the core of my approach is the actor model: millions of lightweight processes that don't share memory, communicate only by messages, and fail in isolation without bringing down the entire node. If you need a service that survives traffic spikes, recovers itself from failures, and holds tens of thousands of concurrent connections on a single machine, this is exactly the runtime I turn to. For years I've thought in terms of let-it-crash, supervisor hierarchies and hot code updates, rather than endless defensive try-catch blocks.
The heart of what I do is properly applied OTP. I build applications around gen_server, gen_statem, supervisors and clean release structure, so that restart strategies, back-pressure and process ownership become explicit, not accidental. I care about soft real-time behavior: predictable latency, bounded message queues and graceful degradation instead of cascading failures. I profile through observer, trace through recon and dbg, tune schedulers and BEAM itself to match the actual shape of your load. Whether it's a message bus, session router, billing pipeline or a telecom signaling component — I make reliability concrete and measurable.
I'm equally comfortable in the wider ecosystem: clustering distributed nodes, connecting Mnesia or external storage, HTTP and WebSocket layer on Cowboy, integration with your services via clearly defined protocols. To be clear: this is specifically Erlang and OTP, not Elixir, Go or Scala; if you have a real problem around concurrency and uptime, I bring tools that were built for it from the start.
— Designing applications and libraries on Erlang/OTP (gen_server, gen_statem, supervisors)
— Fault tolerance: supervisor trees, let-it-crash, hot code updates
— Massive concurrency and soft real-time message passing systems
— Distributed clustering, Mnesia and release building
— HTTP/WebSocket backends on Cowboy and protocol integration
— Profiling, tracing and BEAM tuning (observer, recon, dbg)
The heart of what I do is properly applied OTP. I build applications around gen_server, gen_statem, supervisors and clean release structure, so that restart strategies, back-pressure and process ownership become explicit, not accidental. I care about soft real-time behavior: predictable latency, bounded message queues and graceful degradation instead of cascading failures. I profile through observer, trace through recon and dbg, tune schedulers and BEAM itself to match the actual shape of your load. Whether it's a message bus, session router, billing pipeline or a telecom signaling component — I make reliability concrete and measurable.
I'm equally comfortable in the wider ecosystem: clustering distributed nodes, connecting Mnesia or external storage, HTTP and WebSocket layer on Cowboy, integration with your services via clearly defined protocols. To be clear: this is specifically Erlang and OTP, not Elixir, Go or Scala; if you have a real problem around concurrency and uptime, I bring tools that were built for it from the start.
— Designing applications and libraries on Erlang/OTP (gen_server, gen_statem, supervisors)
— Fault tolerance: supervisor trees, let-it-crash, hot code updates
— Massive concurrency and soft real-time message passing systems
— Distributed clustering, Mnesia and release building
— HTTP/WebSocket backends on Cowboy and protocol integration
— Profiling, tracing and BEAM tuning (observer, recon, dbg)
