Architecture at scale
Microservices, DDD, and event-driven systems that hold up under millions of daily events.
Lead Full Stack Engineer
I build real-time systems
Engineered with .NET and React on cloud-native architecture that stays fast under load.

Lead Full Stack Engineer with 10+ years of experience designing and shipping scalable, high-performance products across startups and enterprise environments. Deep fintech background — signal backtesting, real-time data pipelines, exchange front-ends, and crypto BI — alongside multi-tenant IoT platforms and AI assistant products. Expert in end-to-end delivery with .NET/C#, React/Next.js, and cloud-native architectures on Azure, from event-driven microservices to real-time pipelines processing millions of events a day. A hands-on technical leader who mentors engineers, drives architecture decisions, and pairs DevOps discipline (CI/CD, Docker, Kubernetes) with AI-assisted development to ship faster without sacrificing quality.
Microservices, DDD, and event-driven systems that hold up under millions of daily events.
Leading cross-functional teams, mentoring engineers, and owning the hard trade-off decisions.
Translating business goals into production-ready features that ship fast and stay reliable.
The first thing I do on any product is talk — with coworkers, stakeholders, and customers — until the requirements and the goals behind them are fully understood. Most engineering failures are really understanding failures, and no architecture can rescue a team that is building the wrong thing.
As a lead engineer, the highest-leverage work I do is helping others learn. A team where everyone grows and can see their own impact ships faster, makes better decisions, and is what turns a company into a place people want to build in.
No perfect success stories. What went wrong, what I learned, and what I changed matters more than what worked on the first try — every technical decision is a trade-off, and pretending otherwise is how systems rot.
AI tools like Claude and GitHub Copilot are now part of how software gets delivered, and I use them daily. They amplify engineering judgment rather than replace it: combined with solid system design they compress the path from idea to shipped feature — without that foundation, they only build the wrong thing faster.
The long-form version — real architecture decisions with their costs included — is on the blog.
Lead Full Stack Engineer
Leading a cross-functional team building a multi-tenant SaaS platform on .NET 10 and React.
Senior Full Stack Engineer / Tech Lead
Built a multi-tenant IoT platform end to end; later returned to lead the TDX.bz startup.
Senior Full Stack Engineer / Tech Lead
Built real-time signal-processing infrastructure and led the migration from monolith to event-driven microservices.
Senior Full Stack Engineer
Took Dobord, a gamified cashback platform, from architecture to launch.
Full Stack Engineer
Built RESTful services for distributed systems and grew into microservices and event-driven design.
Mediana — cloud-native backend + React dashboards
A multi-tenant SaaS product behind a YARP gateway: every request authenticates at the identity service, then routes to .NET 10 microservices talking gRPC with per-service Redis caches. Apache Kafka absorbs the high-throughput ingress, SQL Server serves hot data with a ClickHouse cold archive cutting storage costs, and the whole fleet runs highly available on Kubernetes.
IroTeam — devices, apps, and hardware on one platform
An IoT platform managing 5,000+ connected devices with real-time processing of 2M+ daily events: .NET microservices with MQTT ingestion, Nuxt web apps, an Angular admin dashboard, and hardware integrations from greenhouses to access control.
Wallex — signal backtesting & market intelligence
The fintech line of work, with signal backtesting at its core: Kafka and Microsoft Orleans actor pipelines replay the full signal set against point-in-time market data, feeding a ~780-file C# BI backend with a dedicated query API and Nuxt dashboards. The same foundations carried the TradeX BI data warehouse — Orleans aggregation grains behind a Next.js analytics dashboard — and TDX.bz, shortlisted among 35 of 1,043 startups for the Web Summit Qatar 2025 PITCH competition.
Audiobook platform — mobile, dashboard, and backend
The Shahrzad audiobook & podcast platform, built full stack: an ASP.NET Zero backend of 1,500+ C# files with a GraphQL API, the React Native consumer app, and the Angular creator dashboard — install and download analytics, paid-subscriber play-time metrics driving creator revenue share, and financial reports with CSV export.
React Native audiobook & podcast app
A React Native audiobook & podcast app for iOS and Android that I built solo: streaming playback, subscriptions with in-app payment (Charkhone SDK), and catalog discovery over a GraphQL API via Apollo Client with persisted cache — one codebase, both stores.
Video-conferencing platform on ASP.NET Zero
A full video-conferencing product: WebRTC meeting rooms with camera and mic controls, screen share, in-meeting chat, participants, recording, and file sharing — an ASP.NET Zero modular monolith (.NET 5) with an Angular 12 panel, SignalR signaling, a GraphQL endpoint, and Xamarin mobile heads.
Meeting-room reservation & management
An online reservation platform for meeting rooms (Asarad, 2018–2019) built full stack with Angular and .NET: guest invitations with QR access codes, meeting locations with maps and parking info, in-meeting polls and votes, shared meeting files, and notes.
Live online exam-prep classroom
An online entrance-exam class platform I built as primary author (2019–2021, the lms-famiran codebase): an ASP.NET LMS solution with dedicated web, services, and data layers, Video.js live streaming, and an MQTT-powered real-time classroom chat with host moderation, polling, and participant management. Still live at tamland.ir.
Telegram healthcare bot + doctor panel
A healthcare communication platform (Asarad, 2017): patients talk to specialized doctors through a Telegram bot — test-result interpretation, medication questions, preliminary diagnoses — while doctors manage patient lists and respond through a dedicated web panel over a patient-management backend.
Enterprise retail community app — React Native
A community platform delivered for Walmart as an enterprise retail engagement, spanning ~20 repositories. I authored both generations of the React Native app (1,200+ commits across the original MobX-State-Tree codebase and its rewrite, 2021–2022), integrating with enterprise APIs alongside the backend, ML-crawler, and infrastructure teams.
Asarad — gamification startup launch
A gamified cashback platform architected for 1,000+ concurrent users, with MQTT pushing live cashback and leaderboard updates into an Angular front-end, and offer integrations with leading brands including Digikala and SnappFood.
Kafka partitions into an Orleans actor cluster
Event streams land on partitioned Kafka topics and fan out to a Microsoft Orleans cluster where every entity is a virtual actor holding its hot state in memory — no locks, no shared DB contention. Results write through to Redis for millisecond reads and SQL for durable history: 5M+ events a day at millisecond-level latency, scaled by adding partitions and silo nodes.
MQTT uplink, rules engine, command downlink
5,000+ devices authenticate with per-device credentials and publish over ACL-guarded MQTT topics — the topic tree is the API contract. Ingestion workers persist tenant-isolated telemetry, a rules engine raises alerts into per-tenant web and mobile clients, and a command service pushes control messages back down the same broker, so 2M+ daily events and the control path share one secured channel.
A different axis per layer, chosen deliberately
The full scaling map of a production platform serving ~2k requests/sec: stateless API replicas multiply horizontally behind the load balancer, Redis grows vertically first and then splits by keyspace, Kafka scales by partition count chosen up front, and the relational core scales up with a tuned, cache-first read path — sharding stays the last resort, not the default.
AKS namespaces, HPA, first-class telemetry
Production and staging live in separate namespaces with their own config and secrets, deployed from the same manifests; traffic enters through the ingress controller to services fronting HPA-scaled deployments with liveness and readiness probes, and OpenTelemetry ships traces, metrics, and logs from every pod into Splunk and Elasticsearch dashboards — no SSH, no guesswork.
Cloudflare says agents write code faster than teams can review, deploy, and maintain it. I run that lifecycle on a one-person site, and the bottleneck is real.
Read articleA null-placement wrapper in generated SQL made every sorted query a full scan. Why medium data still timed out, and the fix that shipped in Hasura v2.50.0.
Read articleSonar says 42% of committed code is AI-generated or assisted, 96% of developers do not fully trust it, and only 48% always verify. Triage is the new skill.
Read articleAll articles, with full-text search, are on the blog.
Open to lead engineering roles, architecture consulting, and ambitious product work. The fastest way to reach me:
Heads up: the portfolio API isn't running, so the form below will suggest emailing directly.