DistroQ
A high-throughput distributed task queue built in Go with exactly-once delivery semantics. Handles 100k+ messages/sec using Kafka and Redis for deduplication.
About Me
I'm a backend engineer who thrives at the intersection of distributed systems and machine learning. I design APIs that scale, train models that generalise, and write code that doesn't wake me up at 3am. Currently obsessed with high-throughput data pipelines and efficient inference serving.
Skills & Tech Stack
Selected Work
A high-throughput distributed task queue built in Go with exactly-once delivery semantics. Handles 100k+ messages/sec using Kafka and Redis for deduplication.
A lightweight ML model serving framework that wraps Triton Inference Server with a unified REST/gRPC gateway, autoscaling, and A/B testing support.
Real-time anomaly detection for time-series metrics using an LSTM-based model served via a streaming pipeline. Integrates directly with Prometheus and Grafana.
A secrets synchronisation service that propagates Vault secrets to Kubernetes namespaces, AWS Parameter Store, and GCP Secret Manager — with audit logging.
Experience
Acme Corp
Led the redesign of the core payments microservice, cutting p99 latency from 800ms to 45ms. Built an ML-driven fraud detection pipeline that reduced chargebacks by 34%.
DataBridge
Designed and shipped a real-time data ingestion platform processing 5TB/day. Introduced gRPC across internal services and reduced inter-service latency by 60%.
StartupXYZ
Built REST APIs for a B2B SaaS platform serving 200+ enterprise clients. Contributed to the ML feature store that powered personalised recommendations.
Contact
Open to backend roles, ML infra challenges, and interesting problems.
If you have one, drop me a line.