01Live
LLM Foundations
Climb a 12-rung ladder from a memoryless prompt to GraphRAG, evals, agents, and MCP — each step queryable live.
Pythonpydantic-aiOpenAIGraphRAGMCP
Anna Mosaki
Quantitative Researcher · Data Scientist · AI Engineer
Quantitative researcher and data scientist, graduated from ENSAE Paris, with 16 months at BNP Paribas CIB building machine-learning models for trading-desk risk monitoring and NLP analysis of trader communications. Over the past year I built GenAI systems covering RAG/GraphRAG, multi-agent orchestration, MCP and LLM evals.
scikit-learnTensorFlowPyTorchNLPTime SeriesLisbon, Portugal
Builds
Projects
Live demos — open one or dig into the source.
02Live
Agent Desk
Multi-agent investment desk over FastA2A — live graph of agent traffic with human approval gates.
pydantic-aiFastA2AMCPFastAPISSE
03Live
Research Digest
Live ArXiv + fund/quant RSS digest on time series × finance — free sources, SSE regenerate.
ArXivRSSFastAPISSENext.js
Highlights
Award
1st Place
GenAI Hackathon — AWS × Sia Partners × Mistral × Gide
RAG system for legal-document drafting.
Award
1st Place
H-W3B Hackathon — Sia Partners × Tezos
Blockchain-secured digital passport for vehicles.
Award
Best End-of-Studies Finance Internship Prize
ENSAE Paris
Awarded for the BNP Paribas CIB quantitative research internship.
CV
Experience
BNP Paribas CIB
Oct 2024 – Sep 2025
Quantitative Researcher – Data Scientist
Paris, France
- Research and production of ML/DL time-series anomaly-detection models with streaming inference across product families.
- Collaborated with traders to source data and iterate on the product; NLP analysis of trader communications for market-conduct surveillance; two-week training in the London office.
- Selected for the BNP Paribas CIB Graduate Programme in Quantitative Research.
BNP Paribas CIB
Jun 2024 – Sep 2024
Quantitative Researcher – Data Scientist, Summer Intern
Paris, France
- Built and benchmarked deep-learning time-series anomaly-detection models on trading data.
Les Associations Mutuelles Le Conservateur
Jun 2023 – Aug 2023
ALM Modeling Intern
Paris, France
- Modeled portfolio Vega sensitivities in VBA; automated recurring financial simulations for the investment team.
Selected Projects & GenAI Training
- LLM Foundations: 12-level GenAI system — prompts → GraphRAG → evals, agents, MCP (Python, pydantic-ai).
- Agent Desk: multi-agent investment desk (research/macro/quant/risk/scribe) with FastA2A, HITL gates and dual MCP.
- Research Digest: ArXiv + fund/quant RSS desk for time-series × finance research.