Career Transition and Transferable Skills Questions

Situation: I was working as a data analyst at a mid-size e‑commerce startup focused on cohort and A/B analysis. After two years I enjoyed the analytics thinking but frustrated that insights weren’t consistently adopted—stakeholders lacked easy, trusted access to metrics.

Task: I decided to shift into BI to focus on building self‑service dashboards and reliable reporting infrastructure so insights would drive day‑to‑day decisions. My goal was to be credible technically and with stakeholders within six months.

Action:

• Upskilled: completed a Looker/Power BI specialization and an advanced SQL course; learned dimensional modeling and basic ETL patterns.
• Built credibility with proof: led a pilot to replace three manual weekly reports with an automated, documented dashboard (Looker) tied to a canonical metrics layer; implemented row-level security and scheduled extracts.
• Partnered with PM and ops to define SLAs for metric freshness and created a data dictionary + training sessions for users.
• Instrumented usage tracking and iterated on UX based on feedback.

Result: The pilot reduced report preparation time by 60%, increased dashboard adoption to 85% of intended users, and exposed a recurring reporting discrepancy that, when fixed, improved revenue attribution accuracy by ~4%. The success led to a permanent BI role and ownership of the reporting stack.

Why Lyft: Lyft’s scale, real‑time operational data, and emphasis on experimentation align with my BI focus—building reliable, self‑service metrics that enable fast product and ops decisions. My experience designing metric layers, automating ETL, and partnering cross‑functionally maps directly to Lyft’s needs for accurate, scalable dashboards that drive day‑to‑day decisions across rider, driver, and marketplace teams.

Situation: In my current BI Analyst role I build dashboards, automate reports, and surface insights that help leaders make decisions. Over time I found myself wanting to move from describing what happened to building systems that predict and act.

Task: I decided to transition toward AI engineering to apply modeling, automation and product-driven thinking to data problems rather than only producing one-off reports.

Action:

• I taught myself Python, scikit-learn and basic deep learning, built a sales-forecast model that improved accuracy vs. baseline by 18%, and productionized it as a daily API using Flask and scheduled ETL jobs.
• I volunteered to add a predictive layer to our churn dashboard so users could prioritize interventions; I collaborated with engineers to instrument data pipelines and monitoring.
• I focused on MLOps principles (CI/CD for models, data validation, model drift alerts) so models remained reliable in production.

Result / Motivation:

• What excites me most is turning insights into automated decisions and scalable products: feature engineering, model evaluation, observability, and UX that surfaces predictions to users.
• My BI background—understanding stakeholder needs, KPIs, and data quality—gives me an edge: I design models that target real business outcomes and measure impact.
• Long-term, I want to lead AI product teams building predictive and personalization systems that drive measurable metrics (e.g., increase retention by X%, reduce cost-to-serve, improve conversion lift). I aim to own end-to-end ML lifecycle and mentor others in best practices, balancing rigor (accuracy, fairness, latency) with business value.

Why AI vs. reporting:

• Reporting is vital, but I’m drawn to building systems that proactively change outcomes at scale. AI lets me move from hindsight to foresight and automation, leveraging my BI strengths to deliver higher-impact, repeatable solutions.