AI in Finance NYC

Join us at The Rooftop at Club Quarters World Trade Center, a stunning venue in the heart of the Financial District, for an evening of conversation with the researchers, engineers, data scientists, and students in NYC who are working at the intersection of AI, machine learning, and quantitative finance.

No slides. No presentations. No pitches. Just drinks, honest conversation, and meeting the people in this city who care about the same hard problems you do — reinforcement learning in stochastic environments, deep learning for time series, NLP on financial text, backtesting methodology, and everything else at the cutting edge of quantitative AI research

The quantitative AI research community in New York is massive but fragmented. The academics are in their departments. The industry researchers are behind NDAs. The ML engineers are heads-down building. The independent tinkerers are working alone at 2 AM. Rarely do these people end up in the same room.

Come with whatever you're working on, whatever you're stuck on, or whatever paper you read this week that made you rethink something. Bring your questions, your half-baked ideas, and your honest takes on what's actually interesting versus what's just hype.

The Rooftop at Club Quarters World Trade Center has a beautiful lounge and bar with the kind of atmosphere where these conversations happen naturally — sophisticated without being stuffy, relaxed enough to actually talk.

Who should come:

Data scientists, ML engineers, quantitative researchers, financial engineers, PhD students, academics, independent researchers, and anyone in NYC who is serious about AI applied to quantitative problems. All experience levels welcome. Whether you're publishing papers on temporal fusion transformers or just starting to learn PyTorch — you belong here.

What we talk about:

Reinforcement learning architectures for sequential decision-making. LLMs applied to earnings calls, SEC filings, and unstructured financial text. Deep learning for time series — transformers, CNN-BiLSTM, neural forecasting. Factor modeling, signal research, and feature engineering for noisy data. Backtesting methodology and all the ways to fool yourself. Market microstructure and order book modeling. Alternative data as a research subject. Infrastructure — Python, PyTorch, Spark, and the engineering that makes research reproducible. And whatever else comes up when you put a room full of quantitative minds together with drinks in their hands.

🍸 Drinks and food available for individual purchase at the bar and restaurant ⏰ Arrive when you can — the earlier you come, the more people you'll meet

This is not a talk. This is a lab.
Bring your laptop. We're going to build a complete feature engineering pipeline for financial time series data from scratch — pulling market data, computing technical indicators, constructing ML-ready feature matrices, and setting up a proper walk-forward evaluation framework. By the end of the evening, you'll have a working notebook that you can take home and extend into your own research.

What we're building:
We'll work through a hands-on notebook together, step by step, with time to experiment and ask questions at each stage.

Stage 1 — Data acquisition and cleaning Pulling historical equity data using open-source tools. Handling missing values, adjusting for splits and dividends, and structuring raw market data into a clean, research-ready format. We'll build a pipeline that pulls S&P 500 constituent data and constructs a multi-asset universe.

Stage 2 — Technical indicator feature engineering Computing quantitative features from raw price and volume data — moving averages, RSI, MACD, Bollinger Bands, and Ichimoku Kinko Hyo components. We'll discuss each indicator not as a trading signal but as a quantitative transformation of the underlying time series — what information it captures, how it behaves in different regimes, and how to encode it as an ML feature.

Stage 3 — Advanced feature construction Building rolling statistical features (volatility, skew, kurtosis), cross-asset correlation features, momentum and mean-reversion signals, and volume-based indicators. We'll cover how to handle lookback windows, avoid data leakage in feature construction, and think about feature interactions.

Stage 4 — Dimensionality reduction and feature selection Applying PCA to reduce a high-dimensional feature set into orthogonal components. Discussing feature importance, multicollinearity, and how to build a feature matrix that an ML model can actually learn from without overfitting.

Stage 5 — Walk-forward evaluation setup Building a proper temporal train/test split framework. No random shuffling. No lookahead bias. We'll set up a rolling window backtest structure with strict temporal separation between training and evaluation sets — the foundation that every honest quantitative ML study requires.

What you'll leave with:

A complete, working Jupyter notebook that goes from raw data to a clean, ML-ready feature matrix with proper evaluation scaffolding. You can extend this into classification, regression, or reinforcement learning research on your own. The notebook is yours to keep, modify, and build on.

What you need to bring:
💻 A laptop with Python installed (Anaconda or miniconda recommended) 📦 Pre-install these packages before the event: `pandas`, `numpy`, `yfinance`, `ta-lib` or `ta`, `scikit-learn`, `matplotlib`, `jupyter` 🔋 A charger — you'll need the full battery 🧠 Basic Python proficiency — you should be comfortable with pandas DataFrames and basic NumPy. You don't need to be an expert but this is not a "learn Python" workshop.

Setup instructions will be shared in the group discussion one week before the event. Please come with everything installed so we can maximize build time. If you run into issues with setup, post in the discussion and we'll help you troubleshoot before the even

Important note:

This is an educational workshop. We are building research tools and learning quantitative methodology. This is not financial advice and nothing built in this lab constitutes a trading recommendation or investment strategy.