top of page

Cazoo Ltd, London

  • Designed and implemented robust and scalable software systems using Hexagonal Architecture.

  • Provided leadership and guidance to a team of engineers, ensuring successful project execution.

  • Actively collaborated with stakeholders, product owners, and cross-functional teams to align objectives and drive effective communication.

  • Fostered a collaborative and inclusive work environment, promoting knowledge sharing and cross-functional collaboration for innovation and continuous improvement.

  • Played a pivotal role in talent development, offering mentorship and guidance to encourage professional growth and acquisition of new skills.

  • Embraced a strategic mindset, identifying opportunities for process optimization, automation, and adoption of emerging technologies.

  • Evaluated industry trends and leveraged advancements to enhance software development practices and deliver exceptional results.
     

In summary, I combined technical proficiency, leadership skills, collaboration, talent development, and strategic thinking to drive successful software development initiatives and foster a high-performing and innovative engineering team.

Technologies :

​

  • React

  • Typescript

  • MicroFrontend

  • Serverless

  • AWS​

    • DynamoDB​

    • EventBridge

    • Lamda

    • SQS

  • Event Driven Architecture​

  • Introduced Hexagonal Architecture

  • Terraform

  • OpsGenie

Screenshot 2022-09-08 at 18.08.26.png
"Your kindness and support in the workplace has been truly inspiring, thank you for being such an amazing colleague"
"Thank you for your incredible support during my time in the team, I've watched you grow from strength to strength while inspiring others. You brought fun and joy to the team, You always championed needs of others and your empathy is second to none. Continue being awesome and sharing your talent with others on how to lead initiatives successfully"
"Priya is a hard working and driven software engineer. She's very thorough and comprehensive. She's lead several technology initiatives for our team and completed them with high impact. I trust her to do an excellent job and get things done."

Cazoo Journey

Screenshot 2023-03-21 at 16.34.09.png

Carwow

In Q3 ‘22, the Purchasing team need to increase vehicle purchase volume via B2B channels (without impact on other channels / diluting margin to buy more stock). 

​

We decided to join hands with Carwow that allows us to buy 300+ vehicles p/month from customers via Carwow’s concierge service.

​

We already had a journey called Cazoo Purchase to accept cars from Customer directly. However there were some constraints to reuse this journey for Carwow purpose.

The process starts with Carwow sending Cazoo a list of cars that we won bids on. So, we had to come up with a way to accepting this list of cars, giving Carwow customers a link to come to Cazoo and put in their bank details and place an order.

​

I led the discussion for brainstorming a few options on this along with multiple teams and found a smallest possible slice to achieve this goal.

Carwow initiative - Option 4.jpg

Along with leading the engineering discussions for this initiative with multiple teams, I also created few visual representations of this process from both business and tech perspective.

image.png

Learnings

Challenges:

  • ​Thinking from the holistic journey: from Valuations to Handovers

  • Cross team collaboration across different worlds

  • Collaboration with Carwow people

  • Finding the most secured way to transfer file from Carwow to Cazoo

​

Learnings:

  • Efficiently leading technical discussions for two teams

  • Speaking Technical, Business and Product languages in a same session

  • Finding the correct trade-off to achieve the goal

  • Learning Retool
     

Outcome:

  • Data Driven: +£400 GPU p/order on the additional 300 vehicles, approx £120k (400*300) p/month

  • Customer Focused: Smooth process between Carwow and Cazoo till date 

  • Fast Drivers: We got early feedback for the experiment as we sliced it thin

Customer vehicle condition grading

What was the problem?

Screenshot 2022-09-08 at 17.42.49.png

We were seeing some non-ideal customer behaviour in checkout.

​

Namely customers seem to not be seeing the conditions of valuation and are surprised when their booking is canceled or handover is denied and the frequency of customers entering good and excellent condition is much higher than us seeing these conditions.   

How did we solve it?

  • Participated in a lot of user research and user interviews

  • Summarised the mindset of users while selecting car's condition

  • ​Proposed smallest possible experiments to help users to decide car's condition

  • ​Experiment 1: Set default to “Good”

    • Results:

      • Lower percentage of all valuations are with condition excellent (down 3.8%)

      • Lower percentage of all orders (down 5%) - Cazoo Purchase and PartExchange

      • Very big ROI (measurable change with minimal work)

  • Experiment 2: change “Excellent” to be “Like new”​

    • ​Result: After 2 weeks, we managed to see a 4% drop in people selecting the highest vehicle condition when presented with “Like new” instead of “Excellent”. 

  • Experiment 3: Change order to “Worst” to “Best”​

    • ​1% drop in “Like new” selection

Screenshot 2022-09-08 at 17.51.06.png

Experiment 4: Car Condition Quiz

The Problem

Screenshot 2022-09-08 at 17.55.54.png

Inaccurate condition grading by Customers and Handover Specialists was costing us £690,000 per month.

​

Eventually move away from condition categories towards more detailed refurb cost calculation. 

  • More accurate valuation for customers

  • Less risk and higher margin for Cazoo

How did we solve it?

Screenshot 2022-09-08 at 18.03.02.png

  • Participated in user research: what kind of questions users might be comfortable to answer

  • Designed Tiny(slim) MFE architecture for both Valuations(us) and Handovers team to keep all questions, answers and formulas on a single source of truth.

  • Acted as a lead engineering contact for multiple teams

  • Communicated to business, product and technical stakeholders in their own language

Learnings

Challenges:

  • ​Thinking from the holistic journey: from Valuations to Handovers

  • Cross team collaboration across different worlds

​

Learnings:

  • Efficiently leading technical discussions for two teams

  • Speaking Technical, Business and Product languages in a same session

  • Designing architecture whole initiative for both teams
     

Outcome:

  • Data Driven: 28% of customers were in for speedy valuation

  • Customer Focused: 2% drop in Excellent category selection

  • Fast Drivers: We got early feedback for the experiment as we sliced it thin

Part-Exchange for Spain

What was the problem?

Screenshot 2022-09-08 at 16.53.45.png
Screenshot 2022-09-08 at 16.54.57.png

Unlike Part-Exchange journeys for Germany and France, For Spain users, we had to accept VRM instead of VIN.

  • Our entire architecture for EU journey was reliant on car's VIN

  • We had to come up with a way to re-think the architecture design to be acceptable of both car's VIN and VRM

How did we solve it?

image (1).png

  • Introduced configuration based architecture

  • Each market has it's own configurations so that we can scale it to as many countries as we want with all kinds of different features to it.

  • Addressed the riskiest part first, led the development for the configuration based architecture

  • Sliced it really thin to get early user feedback

Learnings

Challenges:
 

  • Architecture

  • Making it accommodate both VIN and VRM in all code bases (valuations-api, customer-vehicle-api, customer-vehicle-ui, revaluation tool).

  • Downstream systems

 

Learnings:
 

  • Identifying opportunity for improvement

  • Negotiation with stakeholders

  • Keeping it simple and stupid

​

Outcome:

  • Customer Focused: Sliced the whole thing into thinnest possible slices. Delivered 1st slice with VIN instead of VRM, but highly scalable architecture

  • Fast Driver: Delivered 1st slice way earlier than expected

  • Team Player: Proactively fostering effective collaboration and relationships with other roles and team

Provenance Checks

What was the problem?

By automating provenance checks, we hoped to:

  • Reduce the amount of Cazoo Purchase cancellations due to failed provenance checks (50% in the last 30 days, 9 Oct - 8 Nov 2021). 

  • Increase handover slot availability.

  • Reduce the workload on our CS teams.

  • Help customers self correct provenance check failures, where appropriate.

How did we solve it?

  • Proposed a solution to bring Provenance Checks at the beginning of the journey

  • Planned the solution presented it to the stakeholders.

  • Designed the architecture and contributed towards developing it from beginning to end

Learnings

Challenges:

  • Communication based on different audiences

  • Data analysis

  • Liaising with Third Party API people 

 

Learnings:

  • Mentoring

  • Design perspective

  • Negotiating

  • Product Observability

  • Stakeholder communication/Written communication
     

Outcome:

  • Customer obsessed: since launch in late November 2021 of just two automated provenance checks we’ve seen:

    • 65% drop in orders cancelled due to failed provenance checks 

    • 45% increase in handover slot availability

    • 32% drop in overall Cazoo Purchase orders (expected).

  • Fast driver: We completed this week one week early 

bottom of page