Анастасія

Anastasiia Martsinkovska
Lviv (Ukraine) • [відкрити контакти](див. вище в блоці «контактна інформація») • [відкрити контакти](див. вище в блоці «контактна інформація») • LinkedIn

Applied AI Engineer with 3 years of commercial software engineering experience building production LLM
applications, data pipelines and automation systems. Strong Python backend background with FastAPI,
PostgreSQL, vector search, APIs, Docker and asynchronous workflows. Experienced in LLM evaluation,
structured extraction, RAG, prompt/context engineering and translating business processes into AI solutions.

Core Skills

Applied AI: LLM Applications, RAG, AI Agents, Agentic Workflows, Prompt & Context Engineering, Tool
Calling, Structured Outputs, LLM Evaluation

Retrieval: Embeddings, Semantic Search, Hybrid Retrieval, Reranking, pgvector, Qdrant, FAISS

Frameworks: LangGraph, LangChain, LlamaIndex, Langfuse, PromptFoo

Backend: Python, FastAPI, Pydantic, AsyncIO, REST APIs, PostgreSQL, SQL, MongoDB

Automation: n8n, Workflow Automation, API orchestration

Engineering: Docker, Git, GitLab CI/CD, Logging, Monitoring, Event-Driven Processing

ML/NLP: PyTorch, HuggingFace Transformers, spaCy, scikit-learn, Pandas, NumPy

Models: OpenAI, Claude, Gemini, Llama

Experience
Unicorn Nest Remote
AI/ML Engineer Jun 2025 – Present

● Designed and deployed a production semantic deduplication service using Python, FastAPI,
PostgreSQL/pgvector, Sentence Transformers and NER, combining embedding similarity with
entity-level signals to identify duplicate articles, achieving 91% precision / 88% recall at scale.

● Built LLM-powered data extraction pipelines integrating Claude/OpenAI APIs with structured
schemas and validation logic to transform unstructured web content into normalized investment datasets
with 84% recall on emerging market signals.

● Built and maintained internal AI/data interfaces supporting human-in-the-loop review, error correction
and downstream enrichment workflows.

● Designed reliable asynchronous and event-driven processing flows with database-backed state, retries,
idempotent processing and structured error handling.

● Integrated AI services with PostgreSQL, internal APIs and downstream processing systems;
investigated production failures across application, database and container layers.

● Containerized services with Docker and supported GitLab CI/CD deployments to DigitalOcean.

Stack: Python, FastAPI, PostgreSQL, pgvector, MongoDB, OpenAI API, Claude API, Sentence-Transformers,
embeddings, LangChain, Named Entity Recognition, Dash, Docker

Synapse AI Remote
Applied AI Developer Dec 2023 - Jun 2025

● Built LLM-powered content workflows using Claude and Gemini APIs, combining structured
prompting, contextual data and automated post-processing.

● Developed an internal RAG assistant with document ingestion, chunking, embeddings and hybrid
retrieval using FAISS / Azure AI Search.

● Implemented semantic entity matching and retrieval logic using embedding similarity and configurable
thresholds.

● Built FastAPI endpoints and Pydantic schemas for LLM-powered services and external integrations.

● Implemented prompt/context optimization strategies to control token usage, latency and API cost,
cutting LLM API costs by ~22%.

● Added structured logging, validation and error handling to containerized AI services.

Stack: Python, FastAPI, PostgreSQL, Gemini API, LLaMA, FAISS, Azure AI Search (vector/hybrid), Docker,
Git, Webhooks, Pydantic, LangChain.

Leadership and Activities
NIKOS LTV Lviv, Ukraine
AI & Automation Instructor Jul 2026
Designed and delivered a practical AI course for non-technical learners, teaching n8n workflow automation,
LLM integrations, APIs, and Google Sheets/Forms through real-world business cases.

Ukrainian Catholic University Lviv, Ukraine
Teaching Assistant in Algorithms and Data Structures Sep 2024 – Dec 2024
Led weekly Algorithms and Data Structures seminars for 50+ students, mentoring students in algorithmic
problem solving and complexity analysis.

Challenger AI, IdeasLab Remote
Startup Accelerator Participant 2023, 2024
Led two AI product prototypes, including an NLP-based disinformation detection system; won UA Recovery
Hackathon and secured a $1,000 accelerator grant.

Education
Ukrainian Catholic University Lviv, Ukraine
Bachelor, IT and Business Analytics 2022-2026

Relevant Courses: Artificial Intelligence, Natural Language Processing, Computer Vision, Probability and
Statistics, Product Analytics, Ethics & AI

Languages: English – B2 (Upper-Intermediate)