Luca Belli Email & Phone Number

Sator Labs offers comprehensive consulting on machine learning strategy and execution, including: algorithm development and auditing, statistical analyses, with a focus on responsible AI and compliance.

Gaithersburg, Md, Us

Timed contract where I was part of the Responsible and Trustworthy AI team.- Promoting and consulting on adoption of the Risk Management Framework in the public (federal, state, city) and private sectors- Co-leading pre-deployment testing and Red Teaming workstreams for the NIST Public Working Group on Generative AI- Supporting NIST’s fulfillment of the Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence by - Developing a companion resource to the NIST AI Risk Management Framework Generative AI - Launching a new initiative to create guidance and benchmarks for evaluating AI capabilities, with a focus on capabilities that could cause harm - Establishing guidelines and processes – except for AI used as a component of a national security system – to enable developers of generative AI, especially dual-use foundation models, to conduct AI red-teaming tests to enable deployment of safe, secure, and trustworthy systems- Representing NIST in ISO Meetings for the 42 series (Artificial Intelligence)

Berkeley, Ca, Us

Part of the inaugural cohort of Tech Policy Fellows, a partnership between the Goldman School of Public Policy and the CITRIS Policy Lab.During my fellowship I was able to further my research on algorithmic amplification for recommender systems (see article below for an example).

Brussels, Be

I was invited to participate in an Expert Group organized together with the Joint Research Center in Sevilla. As experts, we provided a technical report to help the Commission drafting the Delegated Act for Article 40 (data sharing with researchers) of the Digital Services Act.

San Francisco, Ca, Us

Co-Founder and Research Tech Lead, Machine Learning Ethics Transparency and Accountability (META)Conceived and launched Twitter’s first team guiding the application of machine learning practices in responsible and ethical ways. Obtained headcount from executive leadership, grew the team from 2 to 14 in 2 years, and acted as lead and mentor for 3 researchers. Led cross-functional teams to implement concrete product interventions for responsible ML, including: - Quantified and mitigated bias in Named Entity Recognition and Toxicity Model. - Introduced company-wide metric for exposure inequality in A/B tests. - Reduced inequality in the Who-To-Follow user recommendation.Policy work - Published the first ever Twitter political amplification study, delivering insightsinto the personalization algorithm and its effects on US politics and media. - Created algorithmic amplification dashboards in partnership with Policy and Civic Integrity teams. - Shaped Twitter’s internal and external platform-wide recommendation policy.RecSys ChallengeAs Tech Lead for the 2020 and 2021 RecSys Challenges, addressed the lack of large-scale public social network datasets for use by the scientific community when benchmarking new models for tailoring content to user interests. - Developed and released privacy-preserving Tweet and user engagement datasets, each the largest for Twitter at the time. - Released a challenge submission into production. - Introduced the first ever fairness concern for the challenge.

San Francisco, Ca, Us

San Francisco, Ca, Us

Helped jump start a team focused on entity embeddings. We published a blog post on the Twitter official blog and a KDD workshop paper in 2018

San Francisco, California, Us

Remotely teaching Python to both junior and senior developers

Chicago, Illinois, Us

I was a Senior Scientist in the Decision Sciences team, Conversant's Data Science department and directly responsible for overseeing part of the Machine Learning stack that processes millions of pages per month.Some of my projects include:- developing contextual audiences, i.e. ranking users using contextual data- maintaining and enhancing the machine learning stack - image classification using CNNs (implemented with Keras using tensorflow as backend)- an ElasticSearch based recommendation system- developing cross team standards In addition to my technical work, I have taken a leadership role within my team.I was the San Francisco coordinator for our social committee, in charge of organizing fun and team building activities for the group.In the last year I have also taken care of organizing the weekly Lunch & Learn, being responsible for selecting and scheduling speakers and coaching those who need help to get ready for their talks. Under my management, the event attendance has increased and the schedule is much more consistent. I was also a speaker, and I have given the following talks:- Python 2 vs Python 3 (and why we should switch)- From DynamoDB to ElasticSearch: Lessons learned- Deep Q-Learning: teaching your computer to play Atari games- An introduction to functional programming with Haskell- Generating text with Deep Learning and Recurrent Neural Networks (LSTM)- Convolutional Neural Networks for NLP

San Francisco, California, Us

- Mentored and guided students during their 'Data Science Intensive' workshop.

Champaign, Il, Us

- Advanced Research Group member: directly report to the CEO Stephen Wolfram- TextSearch: part of the team that designed and implemented the TextSearch package for Mathematica 10.2 (https://reference.wolfram.com/language/guide/TextSearch.html)- AWS: team member on a project to deploy a deep neural network on Amazon Cloud- MIT IAP Talk: "Intro to Data Science using Mathematica and the Wolfram Language"​- Designed, implemented and maintained the back-end of the Wolfram Problem Generator (http://www.wolframalpha.com/problem-generator/)- Developed continuously deployed revenue generating mathematical features for www.wolframalpha.com

Champaign, Il, Us

- Taught Mathematica course- Responsable for overseeing and mentorning students

Champaign, Il, Us

- Taught Mathematica course- Responsable for overseeing and mentorning students

Rome, Latium, It

Tutored undergrad math classes

Doctor Of Philosophy (Ph.D.), Mathematics

Visiting Ph.D. Student, Mathematics

Master Of Science (M.S.), Mathematics

Erasmus Student, Mathematics

Bachelor Of Science (Bs), Mathematics