Mahdi Naji

Software Engineer, Offering · Snapp!

• Diagnosed a consumer-registration bug that spiked CPU to 100% and then left it idle for extended stretches after rolling deploys; fix cut P99 latency for batch ride-processing from ~5s to ~500ms
• Profiled config processing and caching with Pyroscope, predicted ~12% CPU savings, and shipped a fix that cut CPU and memory usage by ~8%
• Tuned resources across the team's Redis clusters, releasing 60 CPU cores and fixing replica main-thread CPU saturation; fixed a maxmemory misconfiguration (~11% of pod memory) that caused key evictions and was close to a major incident
• Built an LLM-powered ops skill that manages Redis, RabbitMQ and application components, monitors and analyses their health, and records changes, ADRs and future plans
• Working on the core ride-matching engine

Site Reliability Engineer · Snapp!

• Handled incident management and root cause analysis, progressing from on-call shifts into SRE Production Core for the most complex incidents — including reproducing incidents in staging to confirm root cause
• Reduced MTTD to under 6 minutes and MTTR to under 15 minutes
• Hardened super app infrastructure and database platform to support a sister venture's BNPL platform ahead of its Black Friday campaign after prior-year downtime; the campaign ran with zero incidents
• Built a Go library that turns OpenTelemetry spans into unified Prometheus metrics, adopted company-wide, powering a live map of service dependencies and latency
• Built Changall, a full-stack tool for tracking infrastructure changes and managing their approvals, organizing 70+ change requests per month
• Designed business-metrics exporters (SQL databases, Prometheus, ClickHouse) for high-level business metrics and service/domain-specific metrics
• Gathered database metrics into organized dashboards with proper alerting, improving observability across 100+ database machines
• Led incident response for DDoS attacks, with fast detection and mitigation
• Defined SLIs/SLOs for application services across the company
• Built and maintained internal tooling: Prometheus, Grafana, AlertManager, legacy Ansible jobs (deprecated), CI/CD pipelines, LLM-assisted incident-investigation automation, the ELK logging stack, Jaeger/OpenTelemetry tracing, and Pyroscope profiling

Co-founder & Software Engineer · RasaTelecom

• Co-founded an IT services company serving local businesses and factories in Tabriz
• Built VmBox, a Go product integrating WHMCS billing with VMware vCenter to fully automate VPS provisioning
• Owned VmBox end-to-end as a product — marketing, customer support, technical planning, and full product maintenance
• Delivered VoIP calling and automatic-responder systems, virtualization setups, and IoT/Raspberry Pi projects for local businesses and factories

DevOps Engineer · Ramzinex

• Built the monitoring stack from scratch (Prometheus, Grafana, dashboards, alerting)
• Supported migration from async PHP calls to an event-driven, Kafka-based architecture; partitioned topics for exchange-scale load
• Set up GitLab CI/CD pipelines with GitLab Runner-based Docker deployments
• Maintained deployments and monitoring of the PHP/Laravel monolith on the Docker platform; added logging to the code, parsed logs with the ELK stack, and built dashboards for business and technical metrics

Full-stack Developer · Prolar

• Full-stack development on an education platform (course publishing, enrollment, payments) in JavaScript/Meteor
• Built an Elasticsearch-backed search engine for the platform, merging data from multiple sources

Freelance Developer, Tabriz · Freelance

• Built freelance projects for small businesses, including a from-scratch Node.js e-commerce site for a home furnishings retailer
• First hands-on experience deploying and managing Linux servers

Computer Engineering · Tabriz University

Picked up a lot of ideas here, through structured study of important topics like production load, development, and reliability.

Mathematics · Madani SAMPAD High School

Built the electronics for a national neuroscience competition (2nd place in Iran) — a remotely controlled rat maze and a steerable robotic cockroach — which is what got me into C, Arduino, and embedded systems.