Building something exciting! B2B Gen AI.

Specialized in enhancing the safe search experience for over 2 billion users. Developed and fine-tuned machine learning models, integrated them into YouTube's tech stack, and ensured their effectiveness in creating a safer and more user-friendly platform.Key Accomplishments:- Developed and deployed machine learning models, significantly improving the safety and relevance of YouTube's search results.- Implemented a new design and model to robustly defend against harmful and dangerous content, further bolstering YouTube's safety measures.- Collaborated cross-functionally to translate complex model outcomes into tangible product enhancements.- Conducted rigorous testing of models using A/B testing frameworks, leading to a substantial reduction in inappropriate content in search results.- Presented key insights from model analyses to inform strategic decisions and foster continuous improvement in YouTube's safety protocols.

Devised recommendation models to expand sales and usage of Microsoft M365 products. Created and presented data mining and statistical machine learning solutions to address complex business problems. Key Accomplishments:- Built highly effective productionalized machine learning solutions for the deployment of Microsoft Teams.- Contributed to the enhancement of product deployment process by analyzing penetration of Teams Platform for key partners and presenting results to Microsoft Ready (Microsoft's partner’s annual conference).- Successfully filled 13 Patents with the US Patent and Trademark Office, protecting current technologies innovating machine learning space.- Envisioned D&I solution to offer an automatic Name Pronunciation feature to drive a positive working environment (developed during the Hackathon); won the Ireland Hackathon competition and recognized as a top 5 in the US.- Established Causal Inference framework to understand and support drivers of M365 Adoption, as well as presented framework and results to the bi-annual MLADS (Microsoft Machine Learning Conference).

Developed a user behavior predictive model for Microsoft Teams to empower our sales teams. Used Bayesian models, and proficiency clustering.Presenting to the Ready Conference (Microsoft's Partners annual conference), demoing the tool to the sales team in July

Developed and patented a Semantic differentiation clustering method based on the Bayesian Scan statistics to find market-specific issues in NPS feedbacks. (NLP).Patent filed for this project (US 16/444,638)

Built funnel anomaly detection framework supporting teams to promptly uncover any anomalies in conversion funnels (from buying intents to purchase made). Conducted research, analyzed data, and shared insights with peers.Key Accomplishments:- Hosted Microsoft Official Funnel Conference (+300 people) as a Public Speaker, introducing a novel approach to compute "healthy" expected values for the anomalies detected by uplift.- Honored with selecting as a Speaker for opening of new Microsoft Offices (One Microsoft Place) in Dublin.

Created promotional events for CapGemini in my schoolParticipated in training sessions in Paris

Improved processes in the R&D departmentManaged stocks

Repair service