Summary
- Over the last 9 years I’ve worked across data, ML, and software engineering, taking initiatives from vague problem statements to reliable production systems—defining success metrics, designing the architecture, shipping, and iterating based on real usage.
- For the past two years, I spearheaded the development of a B2B SaaS platform called timeshifter that automates billing timesheets for law firms using AI-driven activity monitoring and task linking. In previous roles, I delivered ML and Generative AI capabilities that improved product and pricing decisions, enabled personalization at scale, and supported more tailored customer experiences. That work typically required strong foundations in analytics engineering: clean data models, dependable pipelines, and pragmatic tooling choices that keep systems observable and maintainable.
- Earlier in my career, I interned at Ambev (Craft Beer) and Itaú Unibanco (Revolving Credit).
Professional Experience
- Designed and implemented a full-stack system including a Windows desktop client (C#), Python backend (FastAPI), and PostgreSQL cloud database.
- Designed and built an AI-powered data product that converts raw user activity streams into structured legal timesheet datasets used for billing and analytics.
- Modeled and implemented event-driven data pipelines that ingest application activity, meetings, and document edits and transform them into normalized relational records and task-level timesheets.
- Designed and maintained production PostgreSQL schemas supporting high-volume activity tracking, timeline reconstruction, and timesheet generation.
- Implemented activity clustering, entity recognition algorithms and embedding-based similarity search using vector embeddings to infer clients and matters from historical work.
- Developed websites and dashboards using modern web stacks (React/TypeScript, Next.js)
- Led projects that improved customer experience and increased internal process efficiency using Machine Learning and Generative AI (LLMs).
- Used traditional Machine Learning models and Neural Networks for lead selection and offer optimization. (Keras, Pytorch)
- Data platform engineering (SQL-first): dimensional/relational data modeling and transformations in BigQuery + Microsoft SQL Server, building “clean layers” (e.g., bronze/silver/gold style)
- Owned the design and delivery of data pipelines and algorithms to detect and resolve sales fraud, significantly improving churn outcomes.
- Defined requirements for a B2B CRM platform serving the full customer base, delivered in partnership with a large international engineering team.
- Built multiple predictive models and neural networks to optimize offer strategies, and developed clustering algorithms for customer segmentation.
- Implemented production Python ETL/ELT pipelines, Airflow DAGs, scheduling/backfills, retries, and operational monitoring/alerting
- Specified requirements and developed software to support customer retention operations, delivering individualized offers per customer.
- Implemented a Generalized Linear Model (GLM) in R to identify churn-prone customers and used the model to improve offer strategies.
- Tracked and analyzed churn KPIs to monitor performance and guide iteration.
- Chemical Engineering Laboratory Assistant for Prof. Dr. H. Scott Fogler.
- Assisted in PhD lab research, built their website and assisted in software for modeling Brownian dynamics.
- Analyzed customer journey and market positioning for B2B revolving credit products, influencing team structure and commercial targets.
- Developed statistical models to identify high-potential customers for credit utilization, supporting targeted sales strategies.
- Conducted elasticity tests to optimize pricing structures for credit products, enhancing competitiveness and profitability.
- Developed an accelerated innovation process and management website for the Craft Beer division.
- Created an automated Business Plan model in VBA for new beer concepts, supporting faster decision-making and launch planning.