Professional Experience
Senior Software Engineer, Backend, Agency Handy (SaaS)⁠
08/2023 – present | Dhaka, Bangladesh
  • •Architected the Agency Handy multi-tenant SaaS platform from inception, providing technical leadership to a team of 3 backend engineers.
  • •Designed a high-reliability webhook notification system using BullMQ with per-tenant configurations, exponential backoff retries, and dead-letter queue fallback. Processing thousands of events daily at P99 < 800ms.
  • •Built AWS-native distributed infrastructure: SQS for durable event ingestion, Lambda workers for stateless processing, EventBridge schedulers for deferred job re-queuing, enabling fault-tolerant async workflows at scale.
  • •Engineered a flexible custom field and checkout-flow engine, allowing agencies to deploy bespoke forms and control client on-boarding without engineering involvement.
  • •Integrated Stripe Connect and PayPal Partner for multi-tenant payment processing, supporting sub-merchant onboarding, payout splitting, and secure transactions across hundreds of agency accounts.
  • •Implemented SES/​SNS-backed transactional email with a module-scoped dynamic variable templating engine, enabling branded agency communications at scale.
  • •Led Slack integration enabling real-time workflow notifications and automation hooks, reducing manual overhead for agency teams.
    • 06/2022 – 07/2023 | Chittagong, Bangladesh
  • •Architected a multi-tenant e-commerce platform with custom domain routing (Caddy), delivery integrations (Paperfly, Pathao), and advanced analytics dashboards.
  • •Designed and continuously evolved full database architecture (MySQL) to support new features, third-party integrations, and scale demands.
  • •Led strategic partnership discussions and technical integration initiatives with third-party service providers to ensure alignment with platform objectives. Led the integration with bKash, Paperfly, Pathao, SSLCommerz.
  • •Built a campaign management system enabling merchants to design and execute targeted customer acquisition campaigns with performance tracking.
  • •Implemented automated CI/​CD pipelines via GitHub Actions, enabling safe and continuous delivery across all environments.
    • Education
    BSc. in Computer Science and Engineering, International Islamic University Chittagong
    Chittagong, Bangladesh
    Skills
    Backend: NodeJs, ExpressJs, Fastify, Prisma, Mongoose, Event Scheduler, Cron Jobs, Webhooks
    Authentication: JWT, NextAuth, BetterAuth
    Databases: MySQL, Postgres, SQLite, Redis, MongoDB
    Cloud: AWS EC2, S3, Cloudfront, ALB, SES, SQS, EventBridge, Lambda, SNS, Microsoft Azure, VM, Blob, Docker, Caddy, Nginx, Unix
    Languages: JavaScript, TypeScript, Kotlin, Java, SQL
    Analytics: Mixpanel, Google Analytics, Facebook Pixel, Posthog
    Payment Gateway: Stripe, PayPal, Paddle
    AI: Cursor, Codex, Claude Code, Langchain, Embeddings
    CI/CD: GitHub Actions
    Projects
    Vartiq⁠, Developer Infrastructure for Reliable Webhook Delivery

    • Built a developer-focused webhook infrastructure platform abstracting reliable event delivery, covering retry logic, failure recovery, and delivery guarantees away from application code.

    • Designed a two-path failure recovery system: for delays under 2 minutes, uses SQS Lambda visibility timeout API to delay re-delivery in-place; for longer delays, re-queues via EventBridge Scheduler, minimising overhead while maintaining delivery SLAs.

    • Full pipeline: SDK/​API ingestion → SQS queue → Lambda worker → customer endpoint, with branching for successful acknowledgement vs. failed delivery.

    • Stack: Node.js, TypeScript, AWS SQS, ALB, Lambda, EventBridge, Redis, MongoDB.

    • Load tested the ingestion API across 10–100 concurrent connections: sustained ~192 RPS at peak with p50 latency of 434ms and p99 under 1.7s at 100 connections. Zero errors across all runs. Throughput scaled linearly from ~52 KiB/​s at 10 connections to ~491 KiB/​s at 100.

    • Worker throughput: 133 msg/​sec across 31k requests at 512MB Lambda memory, with concurrency capped at 100 — headroom to scale significantly higher on demand.

    HarvyAI⁠, Write email replies with LLM
  • •Can generate replies with context of the original email.
  • •Customize writing tones. Worked with OpenAI & Claude models.
    • Langchain PDF question bot⁠, Take preparation for exam with AI

    RAG pipeline using LangChain to generate embeddings from academic PDFs and synthesise contextual practice questions.

    Languages
    Bengali, English

    IELTS 7.5

    References
    Rabab Khan Rongon, Sr. Software QA Engineer, Head of Quality Engineering, Tixio Technologies AS
    [email protected], +8801760533424