Karim MD
My name is Karim MD. I am a Technical Lead and Senior Engineer based in Tampere, Finland. My work spans GenAI, cloud-native platforms, AI/ML systems, DevOps/SRE, full-stack engineering, and embedded platforms. I have shipped production systems across Unikie, Technology Innovation Institute, Rivian, Nokia, Huawei, CSIT Finland, and Dcard.
My recent focus is building LLM/RAG systems, AI-powered developer tools, MLOps infrastructure, secure release pipelines, Kubernetes platforms, and observability systems for large-scale production environments. currently based in Tampere.
Email: karimmdmirajul [at] gmail.com
Address: Tampere, Finland
More info: [Google Scholar], [LinkedIn], [GitHub], [2nd-GitHub].
Highlights
- Real-world impact: contributed to search, vehicle, and secure platform systems connected to 100M+ users, 170+ markets, zero-tailpipe EV fleets, and next-generation open-source compute security.
- 2026.04: Completed Harvard Business School Online certificate program in AI for Leaders.
- 2023.10: Led release engineering transformation for Ghaf embedded virtualization framework at Technology Innovation Institute.
- 2023.10: Reduced release cycles from 2 weeks to 2 days with Jenkins, GitOps automation, and reproducible pipelines.
- 2022.11: Designed CI/CD and observability systems for Nokia 5G core infrastructure across 20+ global regions.
- 2021.08: Built OTA firmware delivery and fleet observability infrastructure for 100K+ connected Rivian vehicles.
- 2020.02: Built distributed AI/NLP search and semantic retrieval systems for Huawei Petal Search, part of a global search product reported at 100M+ users and 170+ countries/regions.
- 2020.02: Related Huawei ecosystem impact includes Petal Maps mobility integrations such as Bolt ride-hailing across Europe and Africa.
- 2019.01: Built secure IAM and SSO integrations for healthcare, education, and public sector clients in Finland.
- 2016.01: Developed backend services for Dcard, serving 15M+ monthly unique visitors and 1.5B+ monthly page views.
- Open source: DefectGuard computer vision MLOps, ESP32 edge telemetry firmware, and YKI Swedish/English exam-practice software.
- Open source: Top 5 contributor to Ghaf infrastructure and owner of personal microservices projects.
Work Experience
|
2020—Present: |
Technical Lead, AI & Platform |
|
2023.10—2026: |
Platform Engineer & Release Lead, Ghaf |
|
|
2022.11—2023.10: |
DevOps Engineering Consultant, 5G Infrastructure |
|
2021.08—2022.11: |
Senior Software Engineer, OTA |
|
2020.02—2021.08: |
Software Engineer, AI & NLP Search |
|
2019.01—2020.02: |
Software Engineer |
|
2016.01—2018.12: |
Software Engineer, Backend |
|       Graduated | ||
|
|
2026: |
AI, Leadership — Harvard Business School Online |
|
MSc: |
Mathematics and Computer Science |
Projects & Publications
|
|
Ghaf Infrastructure: Cloud-Native Release Engineering for Embedded Virtualization
Open Source
Top 5 contributor. Built multi-architecture remote builders for x86_64 and aarch64, on-prem test agents with connected hardware devices, monitoring, logging, authentication, Nebula overlay networking, NetHSM hardware signing, OCI registry services, SBOM generation, and secure release automation for embedded virtualization.
[Code][Work GitHub]
|
|
|
Next-Generation Secure Open Source Compute Platform
White Paper / Software Supply-Chain Security
Platform work around provenance, SBOMs, model and artifact signing, reproducible builds, hardware-backed trust, and secure release engineering for cloud-native and embedded systems. The goal is to make open-source compute verifiable from source to deployment.
|
|
|
Zero Trust Autonomous Systems Platform (ZTASP)
Publication / Autonomous Systems Security
A zero-trust platform direction for autonomous and embedded systems: identity-aware workloads, secure boot and provenance, policy-driven communications, signed artifacts, runtime telemetry, and defense-in-depth controls for safety-critical compute.
|
|
|
AI-Powered Developer Tools for Code Review, Release Notes, and Documentation
GenAI Platform
Built and shipped LLM-powered developer tools using PyTorch, LangChain, React, REST APIs, retrieval pipelines, prompt evaluation, repository context indexing, and production observability. Integrated the workflow end-to-end into an internal platform and reduced manual engineering effort by 70%.
[GitHub]
|
YOLOv8 + MLflow
Defect Detection
Monitoring
|
DefectGuard: Computer Vision MLOps for Manufacturing Defect Detection
Open Source / Computer Vision MLOps
End-to-end visual defect detection system combining YOLOv8 training, MLflow experiment tracking and model registry, Great Expectations validation, FastAPI inference, Evidently drift reports, Prefect orchestration, DVC reproducibility, Prometheus/Grafana monitoring, Docker Compose, Nginx, and CI.
[Code]
|
|
|
MLOps Infrastructure on AWS and Kubernetes
Built infrastructure for model training, versioning, deployment, rollbacks, model artifact promotion, and reproducible CI/CD pipelines. Reduced ML release cycles by 60% with Kubernetes, AWS, GitOps, and repeatable release workflows.
|
Edge Telemetry
Sensor Streams
MQTT + RTOS
|
ESP32 Telemetry Firmware with Edge Anomaly Detection
Open Source / Embedded Telemetry
ESP32 firmware that samples DHT22, gas, and MPU6050 vibration signals; runs a FreeRTOS task pipeline; computes adaptive anomaly scores on-device; publishes resilient MQTT telemetry with offline queue and retry; persists thresholds in NVS; and exposes UART operator controls.
[Code]
|
|
|
OTA Firmware Delivery Platform for Connected Vehicles
Developed secure firmware distribution, delta updates, rollback mechanisms, real-time fleet observability, WebSocket telemetry, Kafka event pipelines, AWS SNS/SQS alerting, and dead-letter queue architecture for 100K+ connected vehicles.
[Details]
|
|
|
Petal Search: Mobile-First AI/NLP Search at Global Scale
Global Search Product
Worked on distributed search, NLP query understanding, semantic retrieval, recommendations, personalization, vector search, and grounded response generation. Public Huawei materials report Petal Search reaching 100M+ users around MWC 2022 and supporting 70+ languages across 170+ countries and regions.
|
|
|
Petal Maps Mobility Ecosystem: Bolt Ride-Hailing Integration
Huawei Ecosystem / Maps and Mobility
Huawei ecosystem scale extended beyond search into mobility: Petal Maps integrated Bolt ride-hailing service for users across Europe and Africa, showing the kind of product surface where maps, search, local services, and mobile platform infrastructure converge.
[Article]
|
|
|
Distributed AI/NLP Search and Semantic Retrieval
Built distributed search infrastructure serving 25M+ monthly active users with sub-100ms p99 latency, NLP query understanding, semantic search, recommendations, personalization, vector search, and grounded response generation.
[Scholar]
|
|
|
Rivian Climate Impact: OTA Reliability for Zero-Tailpipe EV Fleets
Rivian / Impact Report
Built OTA firmware delivery and observability systems for connected electric vehicles. Rivian's public impact reporting frames R1T, R1S, and commercial vans as zero-tailpipe emission vehicles, with renewable-energy goals and measurable fleet carbon-reduction work.
|
|
|
AI Leadership, Strategy, and Engineering Execution
Harvard Business School Online / Technical Leadership
Applied AI strategy to engineering delivery: model governance, production readiness, evaluation discipline, platform adoption, release safety, and mentoring engineers across cloud-native, AI/ML, Kubernetes, and embedded Linux systems.
[Info]
|
|
|
YKI-SW: Swedish/English Exam Practice Software
Open Source / React Learning Tool
React/Vite practice platform for YKI-style listening and writing workflows. Includes Swedish and English paper modes, live captions, line-by-line translation, auto-scroll during audio, sticky audio controls, writing prompts, candidate sheets, and B1-focused preparation guidance.
[Code]
|
|
|
gpm-distributed-systems-research
Research-oriented distributed system demonstrating microservices architecture, scalability, inter-service communication, and cloud-native deployment patterns.
|
Technical Skills
Miscellaneous
- Certifications: Deep Learning Nanodegree Foundation, Certified Terraform Associate, CCSP, AWS Solutions Architect, Certified ScrumMaster.
- Publications: Published research on Distributed Systems Optimization, listed on Google Scholar.