About

I'm an AI platform and backend engineer with eight years building production ML systems and distributed services in Python and Go — model-serving platforms, real-time feature stores, and multi-agent AI workflows, owned end to end from architecture to on-call.

At Snapp I founded and scaled an 11-engineer AI-platform organization serving 50M+ users, drove an A/B-tested dispatch rollout that added ~100k rides/day, and built the Go inference gateways and feature store that held latency to sub-50ms p99 across 20M daily inferences. Earlier I re-architected the backend that carried the company from regional service to national leader through ~10× user growth.

Now I'm a PhD researcher at the JD-ICE joint European doctorate (UniGe · UC3M · QMUL), studying Vision-Language Models and hardware-aware inference for real-time autonomous-driving edge deployments.

Languages
PythonGoSQLBash
LLM & Agentic AI
Multi-agentTool-callingRAGGraphRAG
LLM Serving
vLLMOllamaHugging FaceFine-tuningQLoRAMLflowAirflow
VLM & Vision
Vision TransformersVLMFP16 / INT8Edge Inference
Backend
MicroservicesRESTgRPCEvent-drivenRedisKafkaNATS
Infra & Cloud
KubernetesDockerAWSGitLab CI/CDPrometheusGrafana
PhD, Joint Doctorate JD-ICE
UniGe (IT) · UC3M (ES) · QMUL (UK)
2025 – Present
MSc, Software Engineering
IAU, South Tehran Branch
2014 – 2018
BSc, Software Engineering
IAU, Kashan Branch
2008 – 2013