Available for opportunities  ·  Oslo, Norway

AI Engineer
& Data Scientist

Building production-grade AI systems — from LLM-powered platforms and GenAI pipelines to MLOps infrastructure on Azure & AWS. M.S. Applied Computer Science, UiT Arctic University of Norway.

Generative AI Machine Learning Azure Cloud MLOps / DevOps Data Engineering
View Projects Get in Touch
Md Saidul Islam — AI Engineer & Data Scientist
Open to work

Md Saidul Islam

AI Engineer & Data Scientist

Oslo, Norway

13 AI / ML Projects
99.2% Best Model AUC
3+ Years Experience

Who I Am

About Me

I am a Data Engineer and AI Systems Developer specializing in production-ready enterprise solutions, with 3+ years of experience building intelligent, data-driven systems across Energy, Finance, Healthcare, and Maritime sectors.

My work spans Generative AI (fine-tuned LLMs, RAG systems, AI agents), Data Engineering (real-time analytics, ETL pipelines, Databricks Medallion Architecture, dbt), Machine Learning (anomaly detection, computer vision, time series forecasting), and Cloud / MLOps (Azure, AWS, Kubernetes, CI/CD).

Completed my M.S. in Applied Computer Science at UiT The Arctic University of Norway, where I conducted thesis research in collaboration with HHS Robotics — developing LLM-based automated data analysis systems that improved anomaly detection precision by 43.5% and reduced training time by 97.1%.

I have built systems processing 10,000+ concurrent users with 99.9% uptime, achieved 95%+ ML model accuracy, and delivered measurable business impact including $5.9M annual cost savings and $1.4M+ projected revenue across 16 production-ready enterprise applications.

Languages: English (Fluent)  ·  Norwegian (Basic)  ·  Bengali (Native)

16Enterprise Applications
99.4%Best Model AUC
$5.9MAnnual Cost Savings
1,115%Peak ROI Delivered

Education

Sep 2022 – May 2025

M.S. Applied Computer Science  ·  UiT, Norway

AI & Intelligent Agents · Algorithms Design · Systems Programming · Grade: B (Thesis)

Transcript

Sep 2016 – May 2020

B.S. Software Engineering  ·  Daffodil International University

Data Structures · Algorithms · Database Systems · Software Engineering

Transcript Diploma

2016 – 2025

Competitive Programming  ·  Beecrowd / URI Online Judge

146+ challenges solved · 85%+ accuracy · Dynamic programming, graphs, data structures

Work History

Professional Experience

Software Developer Internship

Jobswoop AS  ·  Oslo, Norway

Nov 2025 – Feb 2026

Contributed to a B2B recruitment and job-matching platform, collaborating with product teams and stakeholders to deliver data-driven improvements aligned with business goals through customer-driven product development.

  • Analyzed complex data patterns using PostgreSQL (aggregations, CTEs, window functions) to support business intelligence and platform optimization
  • Created technical documentation for API endpoints, system architecture, and database schemas — ensuring clear communication across development and business teams
  • Participated in agile sprints: planning, code reviews, stakeholder demos, and iterative delivery with continuous user feedback loops
  • Gathered requirements, analyzed user needs, and delivered data-driven feature improvements through cross-functional collaboration
Node.jsTypeScriptNestJS PostgreSQLAWS S3REST APIs Agile / ScrumTechnical Documentation

Software Developer

HHS Robotics Warehouse Automation  ·  Norway

Aug 2025 – Oct 2025

Built a real-time monitoring platform for industrial warehouse automation, processing sensor data from robotic equipment and delivering operational dashboards for warehouse operations teams.

  • Analyzed and interpreted time-series sensor data to identify operational patterns, anomalies, and performance metrics — delivering actionable insights through visualization dashboards
  • Developed automated data validation workflows with statistical quality checks and alerting mechanisms to ensure system reliability across 50+ robots
  • Created comprehensive technical reports documenting system performance, data analysis findings, and operational recommendations for non-technical stakeholders
  • Architected and delivered both monitoring systems with 99.5% uptime and zero critical failures in production
PythonFastAPIpandas PostgreSQLReal-time Data Processing Data VisualizationTechnical Reporting

Research Collaborator M.S. Thesis

HHS Robotics & UiT Arctic University of Norway  ·  Norway

Jan 2025 – May 2025

Conducted Master's thesis research on automated data analysis for industrial sensor systems — working independently on complex AI problems with comprehensive documentation and reproducible results.

  • Developed two novel frameworks: ARMADA (Context-Augmented Anomaly Detection with ensemble ML) achieving 43.5% precision improvement and 97.1% training-time reduction
  • Engineered FACTS (LLM-based document intelligence system) achieving 98% factuality accuracy and 68–78% reduction in documentation search time
  • Developed systematic methodology for analyzing sensor data and operational logs — producing reproducible research with comprehensive technical documentation
  • Deployed across 50+ operational warehouse robots with 99.5% uptime and zero critical failures
PythonOpenAI GPTLangChain Statistical ModelingMLflow Anomaly DetectionResearch Methodology

Technical Stack

Skills & Expertise

Generative AI & LLMs

OpenAI GPT-4oAzure OpenAIAWS Bedrock LangChainRAG PipelinesLlamaIndex FAISS / pgvectorPrompt Engineering Fine-tuningAI Agents

Machine Learning & Deep Learning

PyTorchTensorFlowscikit-learn XGBoost / LightGBMComputer VisionNLP / spaCy Anomaly DetectionTime Series Hugging FaceMLflowOpenCV

Cloud & MLOps

Azure (AKS, Functions, ML)AWS Bedrock DockerKubernetesTerraform / Bicep GitHub ActionsAzure DevOps CI/CD PipelinesPrometheus / Grafana Microsoft Fabric

Data Engineering

Apache KafkaApache AirflowPySpark DatabricksdbtAzure Data Factory ETL / ELT PipelinesGreat Expectations PostgreSQL / TimescaleDBRedis Medallion Architecture

Programming & AI Frameworks

PythonFastAPIStreamlit SQLBash / ShellSPARQL JupyterNumPy / Pandas

Analytics & Visualization

Power BIPlotly / DashStreamlit Pandas / NumPyMatplotlib / Seaborn ELK StackA/B TestingStatistical Modeling

Core Competency Levels

Generative AI & LLM Engineering92%
Machine Learning & Deep Learning90%
Azure Cloud & MLOps88%
Data Engineering (Kafka, Spark, Airflow, dbt)87%
DevOps & Infrastructure (Docker, K8s, Terraform)85%

Portfolio

AI, Data & Cloud Projects

Production-grade systems across Generative AI, Machine Learning, Data Engineering, and Cloud infrastructure.

Academic Work

Research

Master's Thesis  ·  UiT, Arctic University of Norway

Computer Science & Computational Engineering  ·  Completed May 2025  ·  Grade: B

Automated Data Analysis with Large Language Models for Warehouse Robotics Applications

Industry Partner: HHS Robotics AS

Developed two novel AI frameworks — ARMADA (Context-Augmented Anomaly Detection) and FACTS (Factual AI Contextualization & Troubleshooting System) — integrating LLMs with domain-specific processing for automated diagnostics in industrial warehouse robotics.

43.5%Anomaly detection precision improvement
97.1%Reduction in model training time
37.2%Document processing accuracy gain
98%Factuality verification accuracy
40%False-positive rate reduction
68–78%Documentation search time saved
Technical Approach
  • Context-Augmented Anomaly Detection (CAAD) — ensemble of Isolation Forest, One-Class SVM, Local Outlier Factor
  • Fine-tuned GPT models for maintenance recommendation generation
  • Multimodal document processing with layout-aware PDF analysis
  • Mathematical factuality verification using TF-IDF and named entity overlap
  • Deployed across 50+ operational warehouse robots with 99.5% uptime

Recognition

Awards & Achievements

1st Runner-Up 2nd Place
AWS Hackathon November 2025

AWS GenAI Hackathon 2025 — First Runner-Up

Privacy Policy & Terms Analyzer — AI-powered browser extension using AWS Bedrock & Generative AI

Awarded first runner-up for an innovative full-stack GenAI application that automatically analyzes privacy policies, terms & conditions, and cookie agreements using AWS Bedrock (Claude Sonnet 4) — surfacing the 3–5 most harmful clauses for users and enabling contextual RAG-powered chat about any policy document.

Full-Stack End-to-end build
1,536-dim Titan Embeddings
RAG Chat Contextual Q&A
Sig V4 Custom AWS Auth
AI & Cloud Services
  • AWS Bedrock — Claude Sonnet 4 for policy analysis & harmful clause detection
  • Amazon Titan Embeddings (1,536-dimensional vectors) for semantic search
  • AWS OpenSearch Serverless — vector database powering RAG contextual chat
  • AWS DynamoDB — NoSQL storage for analyzed policies with full CRUD
Architecture & Engineering
  • Chrome/Edge browser extension (JavaScript) with custom AWS Signature V4 auth — direct browser-to-AWS API without backend proxy
  • Python FastAPI backend — CORS middleware, service layer architecture (Bedrock, DynamoDB, Vector DB), Pydantic validation
  • boto3 SDK integration across all AWS services
AWS Bedrock Claude Sonnet 4 Titan Embeddings DynamoDB OpenSearch Serverless Python / FastAPI boto3 JavaScript Browser Extension RAG AWS Sig V4

Learning & Credentials

Certifications

Professional development across Azure, Data Engineering, MLOps, and Generative AI — verified credentials from LinkedIn Learning and industry platforms.

Let's Talk

Get in Touch

Open to AI Engineer, Data Scientist, and MLOps roles — full-time or freelance. Based in Oslo, Norway.

Contact Details

Email

saidulislambinalisayed@outlook.com

Phone

+47 462 62 219

Location

Oslo, Norway

LinkedIn GitHub Beecrowd

Send a Message