top of page
DDM_AI_Hero (5).png
Client
Bigeye

As Head of Product & Design, I led the research and product design efforts that brought Bigeye’s Lineage Plus and Dependency-Driven Monitoring to market. These features were designed to help data teams focus on the data that matters, reduce unnecessary monitoring, and improve trust in analytics. By addressing core issues in data lineage and dependency tracking, we set the foundation for AI-driven data quality, enabling organizations to proactively detect and prevent data issues instead of just reacting to them. 

By addressing the foundational challenges of data lineage and dependency tracking, we created a smarter, more efficient monitoring system that set the stage for AI-driven data quality. The result? Faster issue resolution, reduced operational overhead, and stronger trust in enterprise data.

Overview

CLIENT | Bigeye 

ROLE | Head of Product & Design 

PLATFORM | Web 

PRODUCT | Data Lineage and AI-powered Issues

TIMELINE | 12 Months 

  • Lineage Plus Development: 8 Months (Launched 2024)

  • AI-Suggested Incidents & Issues: Q4 2024 - Q1 2025

Responsibilities

Research & User Insights: Led discovery interviews, competitive analysis, and usability testing.

Product Strategy & Roadmap: Aligned features with market needs, ensuring integration with AI-driven data quality.

UX & Design Execution: Designed interactive lineage visualizations, granular search & filtering, and real-time monitoring alerts to improve usability.

Collaboration & Implementation: Worked with engineering and customer teams to refine onboarding and adoption strategies.

Goal

Establish a foundation for AI-powered issue resolution by implementing automated lineage tracking and dependency-driven monitoring, enabling faster incident resolution and enhancing data reliability at scale.

Data-Driven Outcomes

30% - Faster issue resolution for data lineage-related problems

50% - Improvement in data trust and confidence across teams

80% - Engagement rate within 90 days of onboarding

Design Process

By following the six-step Design Thinking process—Empathize, Define, Ideate, Prototype, Test, and Implement—I developed a structured approach to solving key challenges faced by enterprise data teams. This process ensured that our solution was user-centered, technically feasible, and aligned with business goals.

Design_Process.png
Research

EMPATHIZE

QUANTITATIVE

I spent 30+ hours interviewing enterprise customers to dig into their team needs and uncover key challenges. To keep things consistent and reliable, I used a straightforward interview guide that ensured we got comparable insights across participants. Here’s what we found, by the numbers:

  • 72% of participants expressed frustration with the manual effort required to resolve data quality issues and wanted more automation.

  • 60% of teams rely on homegrown solutions (Python scripts, DBT tests) but find them insufficient for detecting anomalies and lineage tracking.​

  • 50% of users reported a lack of visibility into data movement across platforms, making troubleshooting slower, and causing operational inefficiencies.

  • 40% of teams mentioned they don’t currently use a data lineage tool but recognize its importance for faster compliance and debugging issues.

  • 100% of participants who evaluated Bigeye found AI-generated issue summaries valuable, but some teams faced adoption challenges due to their learning curve​.

QUALITATIVE

After talking to teams across multiple companies, some clear patterns emerged. The biggest challenges centered around manual data quality processes, the need for AI-driven issue resolution, and gaps in lineage & metadata visibility. Across the board, users were looking for smarter automation, better insights, and easier ways to track their data.

"We don’t have an enterprise-wide lineage tool yet, but leadership wants end-to-end lineage across SAP, Snowflake, and our analytics stack."
— Sagar Raythatha, Colgate​

"AI-generated summaries would be a game-changer for my team. We often spend too much time investigating raw logs before we even understand the problem."
— Philipp Leufke, Vay​

"We have a triage duty rotation to resolve data issues, but it’s inefficient. I'd love a system that automates issue detection and sends smart alerts to the right teams."
— Chris Boden, Super​​

KEY FINDINGS

These key findings shaped the next phase of design exploration, focusing on three major areas: AI-driven incident resolution, the growing demand for lineage visibility, and the different needs of data engineers vs. business users. Understanding these priorities helped us create smarter, more user-friendly solutions that cater to both technical and non-technical teams.

  • Teams want context-aware alerts that include lineage tracking and affected data assets, reducing Mean Time to Resolution (MTTR).

  • 50% of users lack an end-to-end lineage solution, yet they recognize its importance for debugging, compliance, and governance.

  • Business users need simpler, more visual interfaces that don’t require coding to explore data relationships and troubleshoot issues.

Personas

DEFINE

Market Research

DEFINE

QUANTATIVE

Key challenges in the data lineage marketplace include siloed data

systems, clunky manual processes, and tools that are way too complicated for non-technical users. A lot of lineage platforms still don’t offer real-time monitoring, which means data issues can slip through the cracks before anyone notices. On top of that, business users and data stewards often struggle with these tools because they’re built with engineers in mind—not everyday users—and there’s still a big gap when it comes to no-code solutions.​​

  • Data complexity is growing: 80% of organizations operate in hybrid and multi-cloud environments, increasing the difficulty of tracking data movement.

  • Data trust issues: 70% of business leaders cite concerns about data reliability, leading to a lack of trust in analytics dashboards.

  • Regulatory compliance: Financial and healthcare organizations need better traceability and auditability to meet governance standards.

PROBLEM STATEMENTS

After digging into hours of interviews and market research, it became clear that enterprise teams rely on data in very different ways—and our platform needed to reflect that. In large companies, there’s a huge ecosystem of people making decisions based on data, from executives and analysts to engineers and governance teams.
The product experience naturally splits into two sides. On the top half, you’ve got non-technical business users who need data for decision-making, compliance, and governance. On the bottom half, technical data engineers are managing, transforming, and troubleshooting the data behind the scenes. The diagram we created maps this out, showing how these groups interact and depend on each other. By understanding these different perspectives, we built a platform that works seamlessly for both sides, making data easier to access, manage, and actually use across the company.

“As a business user, I need a clear and intuitive way to track data lineage and understand dependencies without requiring deep technical expertise, so that I can ensure compliance, make informed decisions, and maintain data trust across my organization.”
— Non-Technical Persona

“As a data engineer, I need AI-powered insights and automated lineage tracking to quickly identify the root cause of issues and reduce Mean Time to Resolution (MTTR), so that I can focus on solving problems instead of manually tracing dependencies.”
— Technical Persona

Viability_Feasibility_Desirability
Wireframes

IDEATE

I love sketching wireframes by hand—there’s something about pen and paper that helps me think fast and stay flexible as I explore different concepts.

Sketchbook_1_edited.jpg
Prototype

TEST

We kicked off user testing right on the conference floor at Snowflake and Databricks events, giving us a chance to gather real-time feedback from potential users. Since the sales team was demoing it live, I needed to create a fully interactive prototype that was easy to showcase and explain on the spot.

After the first round of testing, the team reported back on user sentiment and overall usability, giving us valuable insights into what worked well and where improvements were needed. From there, I continued to refine the prototypes and test with customers, making sure every iteration felt more intuitive, seamless, and aligned with user needs.

Usabilty Testing

TEST & LEARN

SCENARIO

  • Starting at the impacted report, traverses up the pipeline to understand at a high level the impacted nodes.

    • Can the user find the problematic lineage node within 10 seconds?

  • Utilize hover states to see the job name, what problem it is having, and the timestamp.

    • Did the user correctly identify the job name and the timestamp?

  • Expand the node. How many jobs are within the Stitch ETL tool?

    • Could the user name the job that failed and trace the connected downstream tables and columns visually?

  • Give the user a chance to play with the prototype.

    • Were they able to discern the difference between healthy objects and objects with issues?

SUCCESS METRICS

By testing the prototype, we were able to quickly validate the overall layout, color choices, and workflow, ensuring everything felt intuitive. The results were a huge success—100% of participants, regardless of their technical skill level, could easily understand the lineage graph. They were able to identify where issues originated downstream and track their impact all the way to the furthest upstream reports. With this confidence, we moved forward with higher-fidelity designs, knowing the core experience was clear and effective for all users.

Visual Recognition
100%
Problem Discovery
~8s
Feedback

IMPLIMENT

After sharing the initial prototype at the conference, I took a step back to refine the designs based on customer feedback and user interviews. One thing that really stood out—100% of the customers interviewed mentioned using "stoplight colors" (red, yellow, green) to indicate the health of their data pipeline. That was a lightbulb moment for me! It made perfect sense and became a key inspiration for the high-fidelity designs. To take it a step further, I dug into Google Maps' UX/UI and pulled in some familiar visual cues to make the experience even more intuitive and easy to navigate.

Google_Inspo (1).png
Design System

IMPLIMENT

With just a few strategically designed components, we were able to map over 70 data connectors and organize their data structures into a clear, simplified lineage graph. The real magic came from these small but powerful components—each packed with layered interactions, smart business logic, and intuitive colors and icons. By breaking data down into just three core nodes—Schema, Table, and Column—we created a clean, flexible system that could scale without limits while keeping things easy to navigate.
To round out the experience, we introduced graph manipulation tools like filtering, expanding, and collapsing nodes, making it effortless to traverse large data pipelines. Navigating complex data relationships became as simple as following directions in Google Maps, giving users a clear path to understanding their data flow at a glance.

bottom of page