Who I am
Call me Sreeman. I grew up on Java, and the smell of build servers at 2 a.m. My career has been a steady march through backend systems that needed to scale, communicate, and not fall over when traffic spiked. The tools have changed, but the core obsession hasn't: how do you design software that feels boring in production and exciting when you're building it?
E-commerce taught me the choreography of high-volume transactions and just how much care you need to take with inventory when people expect next-day shipping. The automotive aftermarket forced me to think in decades, not sprints—parts catalogs never die, and compatibility is a beast. Payments demanded paranoia, cryptography, and the humility to know every failed webhook is a customer left waiting. The food industry, on the other hand, tuned my instincts for latency because your dinner order can't wait on a noisy queue.
AWS came to be a guest during my journey and then became a constant thread through those chapters —Lambda when it made sense, EC2 and S3 when it didn't, and until my love Kibana arrived, a toxic love-hate relationship with CloudWatch logs. These days I'm channeling that experience into AI adoption. I'm less excited about flashy demos and more interested in the boring plumbing that lets an enterprise utilizes a model to make Go Live a peaceful day without losing sleep.
Current focus
My brain is currently looping on a single challenge: making the gap between "cool AI demo" and "AI system quietly running in production" vanish. That means obsessing over data contracts, governance that isn't a blocker, and the glue code that lets models play nicely with legacy systems without demanding a rewrite. Expect plenty of notes on adoption, tools orchestration, evaluation pipelines, and more...
Get in touch
If something here resonates—or if you're wrestling with similar problems—reach out. I'd rather trade ideas than watch them rust.