Production latency work at Kotook, measured in Sentry.
Python · Django · FastAPI · Distributed systems
Senior backend engineer & technical lead.
I’m Mohsen. I build and run Python backends, from data models and APIs to performance, infrastructure, and applied AI.
Working remotely · open to relocating to the EU
An 18× campaign spike with a 0.005% 5xx rate.
Architecture, reviews, mentoring, and delivery across platforms.
Selected experience
Work that had to hold up in the real world.
Apr 2024—NowDubai · Remote
Kotook ↗
Senior Backend Engineer & Technical Lead
I own the backend architecture and end-to-end delivery of a real-estate platform and internal CRM on AWS, from early product discussions to production operations.
- Lead a team of 8 developers while staying hands-on with system design, code review, mentoring, and incident response.
- Built an in-page AI assistant, property recommendations, comparison workflows, and tools that validate and complete project data.
- Cut backend p95 by 61%, p99 by 66%, and homepage API p95 by 95% while remaining stable under nearly 2× normal traffic without additional infrastructure.
- Shipped Android and iOS apps on a shared Kotlin Multiplatform core with native Android and SwiftUI integrations.
Works with Django · PostgreSQL · PostGIS · Redis · AWS · Sentry · Kotlin Multiplatform
Sep 2025—NowFars · Remote
Zooma Outlet ↗
Senior Backend Engineer
I helped replace a legacy store while keeping a live commerce business running—not a clean-room rewrite.
- Replaced the WordPress store with a Django REST backend and Next.js frontend, improving flexibility, load times, SEO, and day-to-day operations.
- Built multi-warehouse inventory, promotions, the full order lifecycle, reporting, and payment, accounting, and logistics integrations for tens of thousands of products and thousands of monthly orders.
- Tuned store APIs to 310 ms p95 and 560 ms p99, with a 0.005% 5xx rate across more than 12 million requests, and kept the platform stable through an 18× campaign spike.
- Helped grow organic clicks by more than 43% and returning users by 18%, while reducing abandoned carts by 13%.
Works with Django REST · PostgreSQL · Next.js · Locust · Prometheus · Grafana · Loki
Jun 2020—Apr 2024Hormozgan · On-site
Faza Zaman Hormozgan
Backend Engineer & IT Manager
I worked across backend engineering, IT, and production support for public-service products in transport, payments, waste collection, and tourism.
- Designed the backend for a UHF RFID indoor asset-tracking system and helped present it at GITEX Global in Dubai.
- Built APIs and data models for municipal services, from source-separated waste collection to levy calculation and payment.
- Proposed and helped integrate Android POS devices with GPS, bank-card, and citizen-card readers for HormozPay, replacing separate hardware and reducing deployment and maintenance overhead.
Works with Django REST · PostgreSQL · Android POS · CI/CD · Linux
Outside the day job
I learn by building the whole thing.
01
Open source · Backend & AI
VectorLens
My attempt to make multimodal search feel like a normal backend service: text and image retrieval, background inference, live progress, authentication, and storage—all in one inspectable project.
Built with FastAPI · PostgreSQL · pgvector · Kafka · Celery · SSE · SigLIP 2
02
Learning product · Web & mobile
Irregular Verbs
A learning product I built end to end: a 244-verb PWA, 5 practice modes, offline support, reminders, and matching Android and iOS experiences.
Built with PWA · Kotlin · SwiftUI · Playwright
Recognition
From refinery sensor data to a national AI final.
Predictive fault detection for industrial cooling equipment
My teammate and I built an end-to-end ML pipeline for the AI Star competition at Persian Gulf Star Oil Company. The system used vibration, sound, and temperature data to detect and classify failures in refinery cooling equipment.
Our model achieved the competition’s highest accuracy score, and our 2-person team finished 2nd overall alongside strong university and industry teams. It was one of those projects where careful data work mattered more than a flashy model.
- Team
- 2 people
- Result
- 2nd nationally
- Signals
- Vibration · sound · temperature
- Work
- Feature engineering · training · optimization
Writing
Notes I wish I’d had earlier.
11 min read
Browse all notes →
What is actually slowing your database down?
↗A practical tour of ACID costs, B-Trees, LSM Trees, compaction, and Bloom filters—and the trade-offs behind each one.
8 min readKafka is fast for surprisingly ordinary reasons
↗Sequential I/O, the page cache, and zero-copy explain most of the story. Here is how they fit together.
A little more context
I care about systems—and the people who have to live with them.
I’ve spent more than 6 years building and operating products across real estate, e-commerce, CRM, and public services. I still enjoy tracing a slow query, but I’m just as interested in making a code review useful or helping an engineer become independent.
I build practical AI when it solves a real product problem: recommendations, chatbots, semantic and visual search, form automation, and industrial fault detection. I’m interested in the useful middle ground between “ignore it” and “put AI everywhere.”
Backend
Python · Django · DRF · FastAPI · PostgreSQL
Systems
Redis · Celery · Kafka · AWS · Docker · CI/CD
Reliability
Sentry · Prometheus · Grafana · Loki · Locust · pytest
Applied AI
RAG · pgvector · model serving · multimodal search
Get in touch
Have a backend problem worth talking about?
I’m open to senior backend and technical-lead roles, especially where reliability, scale, or platform work matters. Remote works; relocation to the EU does too.