|In the summer of 2020, I had the pleasure of interning in the Data & Analytics department at Cisco Systems. Within Data & Analytics, I belonged to two teams: the AI Innovation Lab and the Data User Experience Team.|
PRODUCT OWNER, Cisco Overflow Team
How can we improve the engagement amongst the Data Scientist community at Cisco without hindering (adding work) their workflow? The solution: a StackOverflow clone specific to Cisco Data Science needs. To address this, I formed an intern engineering team and proceeded to begin design.
On day one, we formed a beta user group of Data Scientists across Cisco, with a fair sample of tenure, business focus, and toolkit. This beta group was involved throughout the design and development cycles.
GROWTH THROUGH DATA
We learned Data Scientists used the specific chat rooms to post & answer questions. I built a data pipeline to scrape all questions and corresponding answers, then auto-populated Cisco Overflow. This provided the platform with hundreds of uploaded questions and answers, solving the chicken-egg issue.
Product Analytics Stack
KPI THOUGHT PROCESS:
What information can we gather if all Data Scientists at Cisco were simply registered on this platform? What executive metrics can be mined? What decision can be made with this data?
AI Innovation Lab
At the Innovation Lab, we looked at possible pilots we could develop using newly cleansed and processed data silo’s. I designed a handful of machine-learning powered products, completing user interface designs and product development requirements (architecture).
What kind of data do we have (and associated metadata) that no one else does? What kind of functionality can we build with a model trained on this data? Is this better than any non-machine learning alternative? How can we implement this without changing the user’s workflow?