Sripati
Sripati's blog

Sripati's blog

It's not a claim, but a try to prove!! Technical Product Manager, Always a great and funfilled Cruise!

Sripati's photo
Sripati
·Mar 30, 2022·

4 min read

It's not a claim, but a try to prove!!
Technical Product Manager, Always a great and funfilled Cruise!

I DON’T KNOW WHO YOU ARE, OR WHAT YOU WANT… BUT WE’LL HAVE A WORKING PROTOTYPE NEXT SPRINT!!

I left my last notion talking about If I am a data product manager?? Finding the answer with the findings I have…..

sripatimishra.in/if-i-am-a-data-product-man..

So, in continuation of what has been said, I continue talking technical product manager's role in creating data products.

What is the underlying problem? Redeveloping a machine learning pipeline by replacing a third-party tool that is impenetrable, it is something there in the solution pipeline, but nobody knows what’s happening behind the curtain of analytics solutions. Development of a platform for disengaging the AutoML tool and setting up ML model functions of user’s preference based on business understanding and requirement for SKU recommendation system.

As said, we developed an end-to-end solution considering all the below factors:

  • Goal of the customer journey on the current solution
  • Empathy Map
  • Persona
  • Customer profile & Value map
  • Facets of Value of the solution
  • Key MVP with Hypothesis

ML1.png Picture by Author

Customer Interview Questions – Data Scientist.

  • Two senior data scientists were interviewed.

Top questions asked to Data Scientist personas:

  • How much is the current solution Scalable and dependable?
  • Question on efficiency and maintainability of current solution?
  • How is the alerting system has been configured?
  • Question on the source of data, data preparation?
  • What is the environment setup?
  • How is the training and prediction model currently configured?
  • How do we track the development of the ML model?
  • How do we deploy the model into production?
  • How are we choosing the best performing Model in the system?

ml5.png

Picture by Author

Data Scientist\Business Analyst Persona

Behavior - With the current solution data scientist and the Business team feel that they are not in decent shape to make decisions for SKU recommendations with the AutoML tool in place in the ML pipeline, keeping them in a space where they are unaware of the ML and data science practices happening behind the automated system, leading to a situation where data science team along with business leaders are having a constricted knot on their hands to tailor the data feature, performing hyperparameter tuning, tuning the solution, understanding the training paradigm of ML model as everything appears to be a black spot in the pipeline.

ML3.png

Bio: Mr. Bhatt is a business owner for SKU unit in the company. He is aware about the ML solution we have in the system for recommendation, but when it comes to comparing the result and performance of solution it is having low precision and recall.

ML4.pngPicture by Author

MVP-Hypothesis

Hypothesis statement : " Creating a solution that is certainly available, controllable, and accessible. One can improve the handling of unstructured data and transform it into a meaningful asset for the company. Getting into a space where best user experience adherence can be found."

The above research indicated that even though we have a surfeit of data science practices and solutions available in the market, still the personas are not getting the destined solution that they are looking for bringing the best insight into the business developments. AI/ML is the disruptive fulcrum of modern days working processes, where communities are willing to make the best use of this technique for growing business, having better customer coverage, and holding stiff gravity in the market.

I tried to keep myself limited to data science's mathematical impact on the solution that has been implemented in this product. Being a product manager, one should be prescient in highlighting the right trade-off of developing product vision, value creation, and keeping empathy for customers whenever one talks product.

DATA PRODUCT.

Areas of ML solution which further I can talk about being a product manager will be

  • Data product design Thinking
  • Empathizing with Data scientist, data engineer, and business analyst
  • How to define the problem
  • Ideation of solution for data product
  • Creating MVP
  • Evaluating the solution And likewise, there are so many subjects available to organise with real time execution and product development.

Frankly, all I am trying here to establish a wonderful association between my project experiences in data product and product management theories being put forth by veterans from all over the systems and communities. Its a way to get into the arena of ethical entrepreneurship and contributing to bring world closure to self-objectified experience which one achieves with time.

So, ways to go, next time will talk on something modernistic screed incarnating TECHNICAL PRODUCT MANAGEMENT

 
Share this