Jacopo Tagliabue

you-dont-need-a-bigger-boat. An end-to-end implementation of intent prediction with Metaflow and other cool tools

MLSys-NYU-2022. Slides, scripts and materials for the Machine Learning in Finance Course at NYU Tandon, 2022

recs-at-resonable-scale. Recommendations at "Reasonable Scale": joining dataOps with recSys through dbt, Merlin and Metaflow

post-modern-stack. Joining the modern data stack with the modern ML stack

foundation-models-for-dbt-entity-matching. Playground for using large language models into the Modern Data Stack for entity matching

FREE_7773. Materials for my 2021 NYU class on NLP and ML Systems (Master of Engineering).

paas-data-ingestion. Ingesting data with Pulumi, AWS lambdas and Snowflake in a scalable, fully replayable manner

tensorflow_to_lambda_serverless. Serve tensorflow models prediction from AWS lambda endpoints

no-ops-machine-learning. A PaaS End-to-End ML Setup with Metaflow, Serverless and SageMaker.

dag-card-is-the-new-model-card. Template-based generation of DAG cards from Metaflow classes, inspired by Google cards for machine learning models.

retail-personalization-workshop. In-Session Personalization Workshop for eCommerce, April 2021, and the MICES Workshop in June 2021.

MLSys-NYU-2023. Slides, scripts and materials for the Machine Learning in Finance course at NYU Tandon, 2023.

anki-drive-python-sdk. Python+node wrapper to read/send message from/to Anki Overdrive bluetooth vehicles.

clothes-in-space. Personalization with deep learning in 100 lines of code

pixel_from_lambda. Serve a 1x1 GIF pixel from an AWS lambda-powered endpoint

spark_tree2lambda. Python micro-service to serve a decision tree trained with Spark through AWS Lambda

intro-to-ai-agents-columbia-2025. Code snippets for my 2025 lecture at Columbia University, an introduction to AI Agents

vibe-proving-with-llms. Training an LLM to generate mathematical proofs with a formal verifier in Python

session-path. SessionPath is a deep learning model that provides personalized category suggestions for type-ahead APIs. This repo re-implements the original paper (https://arxiv.org/abs/2005.12781) leveraging Ludwig capabilities.

tarski-2.0. Old-style computational semantics at the time of Python 3.6

LLMs-to-Alloy. Example of LLM generated Alloy code for deductive reasoning from English descriptions.

webppl_to_lambda_serverless. Deploying a webppl probabilistic program as an (AWS lambda) endpoint.

magic-the-gpthering. Playground for generating cards in the style of "Magic The Gathering" using generative AI

On-the-plurality-of-graphs. WIP code for the "on the plurality of graphs" paper

how-much-is-a-billion. Generating meaningful perspectives with NLP and Probabilistic Programming.

jacopotagliabue.github.io. Personal website