Kaushik Mukherjee

I start off by calling out where I am coming from (“in my experience” or “this is what had happened in the past” etc ) that sort of sets the stage. I am trying to say that I am happy to stand corrected in case my inferences are not objective because obviously my experiences aren’t all encompassing. I also sometimes understand my preferences and call out my bias explicitly so people are aware. For eg : I have a preference towards this font style or this particular colour scheme and maybe that’s why I like this version of the design.. I then nudge for opinions. Some align with my thought process, some don’t - that helps triangulate. In the absence of a group I go with knowing there are biases at play, and keep checking frequently.

In my experience a project lacks strategic thinking because of some unwritten rules. Strategy for some strange reason is assigned to “some person” who is also tasked with having the vision and how to realise it. Conversely consider building a framework where strategy is more coherent. Get the team to also agree on the vision. Maybe contribute to the vision. And then ask, how do we get there ? This method ingrains strategy IMO. It nudges individuals to look beyond their immediate responsibilities and ask the “what if” question. What if I did not do x,y,z would it land us closer to our vision or further from it.

Demand sensing and shaping not only has to take Into consideration eventual supply but good shaping techniques also take into consideration existing supply and fulfilment capabilities This works beautifully when done in conjunction leading to significant optimisation in operational costs. What is required to achieve this though are products that can 1. Look at current and past data and predict the likely future 2. Influence the demand 3. Monitor anomalies Platforms like “Percept Insight” are specifically built for this purpose. They have strong observability and anomaly detection capabilities and can influence both internal supply management and demand. Also strong predictable sourcing partners goes is imp towards supply match.

1. Has a sound understanding of the internal functions of any given technology 2. Has an eye towards latest advancements in technology and the wisdom to adapt what might be relevant and reject what might be merely "cool to use" 3. Is able to set expectations clearly and a method towards ensuring predictability in deliveries 4. Is able to collaborate internally as well as cross org towards aligning individuals to the overall organisational goals. Staff engineers typically own contracts between systems 5. Has a sound domain understanding and is able to articulate the trade offs of a given approach. 6. Contributes to the growth of tech quotient of an organisation, both in terms of people as well as in terms of technological artefacts.

1. Please do not be a code monkey. 2. Please understand why something is being asked. 3. Ask clarifying questions to make sure that blindspots are uncovered 4. Contribute to making the requirements solid 5. Your code can have impact on neighbouring functionalities so please consider them. Its better to clarify all assumptions 6. If a requirement smells like a point fix call that out 7. Have a very clear discussion on what the "done" criteria needs to be.

While missing values can occur because of various reasons like bugs in code, incorrect instrumentation, incorrect interpretation and the likes the biggest challenge IMO is network partitions, which can and will occur. The reason missing values are important is because data systems rely on those values for various kinds of statistical analysis.The most important amongst them would be correlation analysis and to a lesser extent factorial decompositions. But the impact of these can be disastrous in terms of false positives or worse false negatives. False negatives are esp dangerous in cases where you rely on models to tell you if an outcome can be expected and the model ignores an obvious outcome .