Abdur R.

Software Engineer, Machine Learning Infrastructure

Koblenz, Germany

Experience

Jan 2024 - Apr 2025
1 year 4 months
Redmond, United States

Software Engineer, Machine Learning Infrastructure

Retrocausal

  • Built a WebSocket progressive streaming pipeline replacing 15+ second blocking fetches with near real time updates.
  • Deployed Nginx with HLS/RTMP for live and on-demand video: reliability up 35%, latency down 30%, concurrency up 45%, bandwidth down 20%.
  • Added client-side caching with AWS S3 pre-signed URLs; image loads dropped from about 5,000 ms to under 20 ms, boosting page responsiveness.
  • Optimized video frame annotation pipeline by introducing Python threading to parallelize S3 uploads, reducing processing time by ~40%.
Jul 2022 - Jan 2024
1 year 7 months
Karachi, Pakistan

Analyst Software Engineer

ibex.

  • Built a custom URL shortener generating over 1 million unique codes per minute with collision resistance.
  • Implemented WhatsApp messaging via Twilio Content API with approvals and delivery/read webhooks; reduced invitation costs by 80%.
  • Migrated APIs from .NET Core 3/5 to .NET 8 and standardized DI/polymorphism/versioning, yielding about 20% faster endpoints and 40% better maintainability.
  • Drove test-first development with xUnit and CI, cutting release lead time from about 4 hours to under 60 minutes and improving stability.
Jun 2021 - Jul 2022
1 year 2 months
London, United Kingdom

Full Stack Developer

Virtual Sandbox

  • Designed and built DulyPaid in Laravel: billing/invoicing automation with payment schedules, dunning reminders, reporting, and analytics.
  • Integrated five accounting systems (Xero, FreshBooks, Sage, FreeAgent, plus one) via REST APIs, webhooks, and scheduled cron syncs; added messaging via Twilio/Nexmo.
  • Set up Jenkins CI/CD with unit and integration test gates to standardize build, test, and deploy and accelerate releases.
  • Implemented Redis caching and connection pooling for high availability and low-latency data access.

Built a CNN-powered GI Tract Disease Classification diagnostic system

  • Accuracy: 92% using PyTorch and Django.
  • Designed preprocessing and automated evaluation pipelines.
  • Integrated real-time dashboards for visualization and collaboration.

Summary

Results-oriented Full Stack Engineer with expertise in scalable systems, cloud-native development, and AI/ML integration.

Proven track record of delivering high-performance, secure applications, including NLP-driven sentiment analysis, customer-support chatbots, and API-first data pipelines built with REST, webhooks, and automation.

Skilled at driving measurable improvements in performance, reliability, and team productivity.

Languages

English
Advanced
German
Elementary

Education

Oct 2025 - Present

University of Koblenz

Master in Web & Data Science · Web & Data Science · Koblenz, Germany

Oct 2018 - Jun 2022

National University of Computer & Emerging Sciences

Bachelor in Computer Science · Computer Science · Pakistan

Need a freelancer? Find your match in seconds.
Try FRATCH GPT
More actions