| Tell me something about yourself |2018-11-20 by Amit
* This is a live document and some of the questions / answers may be updated to reflect my changing opinions. Facts will stay unchanged. *
Tell me something about yourself
I am really into sci-fi movies and have watched all of the good ones available there have been (80s onwards). Also, these days I am totally loving reading articles, papers and theories about machine learning, especially the transfer learning, ontology, unsupervised learning and Bayesian as well as active inference. But I am really shaky with all the maths.
Oh, and you can ask talk to me for hours about privacy, smart-phones, free software, photography and importance of narratives.
The boring answer (may be the one you wanted ?):
- Software engineer turned product management professional with total of 10 years experience
- Initial 4 years working at Infosys for Investment Banking clients; Next 3 years with IGT (working for SITA) and ADP
- Past couple of years as founding partner of own venture, working on NLP based custom solutions for qualitative research
- Strongly opinionated on some topics and biased for data-supported claims / decisions
- Keywords: Roadmap, User Stories, Prioritization, Ideation, Customer Engagement, Scrum, Agile, JIRA, Rally, Mock ups, Balsamiq, Python, Machine Learning, SQL, Shell
What are you doing these days? Anything cool?
Glad you asked!
I am currently working on couple of NLP projects. One of the projects is for improving document clustering in n-dimensional space based on configurable parameters.
In simple words, I am trying to come up with a system to group documents / text in clusters and rank each one in the corresponding cluster based on given criteria (say, grouped by reading-difficulty level and sorted by document size)
Ultimately though, I want to be able to come up with robust systems to
- Semantically simplification of complex documents (e.g, user agreements to bulleted point summary)
- Extraction of well defined (SMART) goals from a natural language job requirement (which may lead to an even larger project for later)
But remember Maths being my weakness when it comes to hard core ML ?
Well, it turns out I am not particularly strong at programming either after being away from day-to-day hands-on coding for past 6 years.
I can still read and understand the code and do some exploratory analysis in Python and hack together some scripts in Shell or may be even JS, but that's about it.
So, another thing I am doing is hacking away at publicly available datasets using Python and in the process exploring SciKit, Pandas, Numpy, MatPlotLib and interesting GitHub projects related to NLP.
Employment related data sets and govt datasets on education are my favorites. Also, I just discovered FastText by Facebook last week (Early November 2018) or so.
Finally, I am also looking for fun teams to work with who are working on interesting problems related to these areas.
Or solve something cool while still being able to remain in touch with these areas ("these areas" being data-exploration and analysis, machine learning models, and other stuff mentioned in the Short Answer about myself).
For more details, have a look at [what I have been up to recently](https://proxygeek.github.io/projects/)
That's interesting! What about some of your previous work - anything I would know or might have seen / used?
Well, my early experience (pre-2010) was with large Investment Banks (with names sounding like Oldman Wax, etc) where I worked on systems flagging anomalous financial transactions.
So, I hope you would not have come across this bit of my previous work.
Post 2010, my work was largely around border security systems. The visa processing, passenger manifest systems and the likes.
You might have had knowingly or unknowingly an experience with some bit of my work in this phase if you have traveled to Middle East, few countries in South East Asia or Australia.
But not something you would have noticed; it was all behind the scenes.
In the off-chance you did come face to face with the "back-end" system, I am really sorry for the inconvenience.
Most recently, in 2015-16, I got to deliver a pretty slick looking Reporting system for HCM (Human Capital Management) folks.
If you are on the payrolls of a USA company, there is about 17% chance that your HR person would have had experience with this shiny new tool.
But do remember: HR people use a plethora of products from different vendors even for related tasks (vendor-diversification and all that). So, if you hear some hate from them for the systems they use, it probably is not ours. Ours was awarded one of the best design refreshes in the industry.
Last couple of years (2017-18) was spent primarily working on some nifty, custom solutions for name de-duping, auto-labeling and categorization of survey-responses, etc. If you have participated in or conducted some such surveys over the past couple of years, there is a non-zero chance of your responses having being processed by our systems. Please note that all such solutions provided were completely owned and operated by our clients and we have no access to any bit of the data processed.
Some questions which usually come up in a job-interview setting
We are big fans of Agile. Do you have any experience working in an agile environment?
I have been working with companies who followed some versions of Agile (Scrum) methodology in one form or another. One of them called it "Factory Agile" while at another place it was just Agile.
In my limited experience over past 6 years, people mean different things when they talk about working in Agile environments.
If you ask about Scrum teams, its rituals of sprint planning, daily stand-ups, reviews and retrospectives - yes, I am intimately familiar with the same.
I agree with the intent behind the methodology - which I believe is to reduce the feedback cycle on the product being built and thereby reduce the possibility of expectation mismatch.
But as often as not, I have seen people following the rituals without really caring for the underlying intent for the same.
In cases like this, scrum / agile becomes more of a religion than anything else - which does not serve the purpose of helping create a good product - delivered as expected.
What excites you about product management?
Nothing. Not specifically about product management.
This question comes up every now and then, especially during some job interviews. What excites me is solving an interesting problem which helps people in a measurable way. If I can help people achieve something in 2 steps instead of the 7 steps they have to take currently, that excites me. Or enabling them to be able to do something they could not earlier, that excites me as well.
Such improvements in the quality of life can be brought about through many ways - including financial, policy or technological intervention. So far, my expertise and experience has been in technology and I am happy to take the technological route for this change.
Product management is just one aspect and one (loosely defined) way of helping solve such problems through technology To that end, I don't mind getting down to hacking something together myself, if required.