National Competitor | Top-tier AI Challenge

• Challenges Solved:
  • Astar Island Oracle: Probabilistic engine for stochastic civilization simulation.
  • Ledger Logic Agent: Autonomous agent for financial/accounting automation (Tripletex).
  • Retail Eye Detector: Computer Vision system for real-time product recognition (NorgesGruppen).
• Tech Stack: Probabilistic ML, LLM Agents, Computer Vision, Python 3.12, Docker.
• GitHub Repositories

Enterprise-Grade Predictive & Generative System

• Architecture: Multi-stage pipeline (Explore -> Engineer -> Predict -> Intervene).
• Innovation: Combines XGBoost for risk prediction with Google Gemini for automated, personalized student interventions.
• MLOps: Production-ready code with 14-day observation windows, hyperparameter optimization, and class imbalance handling.
• Tech Stack: Python, XGBoost, Google Generative AI, DuckDB, MLflow.
• GitHub Repository

Google Cloud/Tekna Hackathon Winner

• Architecture: Vertex AI Agent Framework with Google ADK integration
• Agent Capabilities: Autonomous PR review, multi-agent orchestration, tool-calling implementation
• Tech Stack: Google Cloud Platform, Vertex AI, Python, Microservices, GitHub Actions CI/CD
• Impact: 80% reduction in code review time through intelligent automation
• Live Demo | GitHub Repository

• Built MCP servers for ML experiment tracking and dataset exploration
• Created tools for MLOps teams to interact with Databricks via AI agents
• Documented installation and deployment guides for Windows/macOS
• Technologies: Python, TypeScript, MCP Protocol, Databricks Integration

• Architecture: Multi-modal CNN with FastAPI backend, MCP integration for model serving
• Performance: <100ms inference latency, distributed training pipeline
• Infrastructure: Docker-based microservices, GitHub Actions CI/CD, Azure DevOps deployment
• Tech Stack: PyTorch, FastAPI, Docker Compose, Google Cloud Storage
• API Repository | Web Interface

• Developed comprehensive pipeline for tumor immune evasion analysis
• Integrated RNA-seq expression data with HLA-peptide binding predictions
• Implemented expression-based filtering and weighting strategies (TPM)
• Technologies: Python, Dask, PyTorch, Bioinformatics libraries

• Scale: 1M+ documents with semantic search capabilities
• Agent Framework: Custom MCP server for document retrieval and processing
• Vector Store: Optimized embeddings with multiple retrieval strategies
• LLM Integration: Fine-tuned models with LoRA/QLoRA for domain adaptation

• Agent Frameworks: LangChain, CrewAI, AutoGen, Google ADK, Azure SDK for Agents, LangGraph
• MCP Servers: Built custom Model Context Protocol servers for ML experiment tracking
• Fine-tuning: Production experience with Llama 2, Mistral, and domain-specific models
• RAG Systems: Implemented production pipelines with hybrid search and re-ranking
• Prompt Engineering: Advanced techniques for multi-step reasoning and tool use

• Pipeline Automation: GitHub Actions, Azure DevOps, GitLab CI, Jenkins
• Model Deployment: Vertex AI endpoints, Azure ML, Kubernetes-based serving
• Monitoring: Prometheus, Grafana, custom dashboards for model drift detection
• Infrastructure as Code: Terraform, Helm charts for reproducible deployments
• Experiment Tracking: MLflow, Weights & Biases, Custom MCP integrations

MLOps & Data Scientist | Riverty (Dec 2022 - Present)

• Architected production ML pipelines with automated CI/CD using Azure DevOps
• Built MCP servers for ML experiment tracking and dataset exploration
• Deployed agent-based systems for automated model evaluation and monitoring
• Implemented RAG solutions for internal documentation and knowledge management
• Managed Databricks clusters and optimized Spark pipelines for cost efficiency

Data Scientist | Outshifter (Apr 2022 - Aug 2022)

• Developed cloud-native ML solutions with GitHub Actions automation
• Integrated LLM capabilities for product recommendation systems
• Built real-time inference pipelines with sub-second latency

I'm actively seeking AI Engineer, MLOps Engineer, or Agent Developer positions where I can leverage my expertise in:

• Building autonomous AI agents with LangChain, CrewAI, AutoGen, and MCP
• Implementing production RAG systems and LLM fine-tuning pipelines
• Designing CI/CD pipelines with GitHub Actions and Azure DevOps
• Creating scalable ML infrastructure on GCP (Vertex AI) and Azure

Contact: alenacivanovaa@gmail.com | Location: Oslo, Norway