Related Posts
How much was your bonus this year?
Recs for sushi grade tuna to buy pls? Thanks!
Hey Fishes,
Please help me select from below offers :
1. Capgemini: 18 LPA(5% variable)
2. Qburst: 20 LPA( All fixed)
3. TechMatrix: 21LPA + 1 JB (10% variable)
Yoe : 6
Role : Salesforce Developer
CCTC : 10 LPA
LWD : 31st October
I am looking for other offers also but these are what I have as of now.
Infosys Tata Consultancy Deloitte Accenture
McKinsey & Company A Business Analyst with 2+ years of experience. Looking for job opportunities abroad in McKinsey & Company Skills include data research, analysis, visualization, writing PoVs, primary research (interview &surveys), market analysis etc.
Open to Canada, Australia, UK, Switz and flexible with other places too.
Please let me know if anyone can refer me! Would like to help you out in someway too :)
Additional Posts in Data & Analytics Consultants
Thought this was interesting. Across 160 teams of researchers, just about all failed to make good life outcome predictions on things like GPA, evictions, layoffs, and others. Data followed 4.5k families across 15 years, with 13k features (varied over time). Haven't looked at it directly yet, but will be turning the docs and data inside out... In the meantime, authors claim this as showing the limits of ML. Oh, and it's published in PNAS, so you know there's some big publication energy there.
https://www.pnas.org/content/117/15/8398
What is a data lake in basic terms?
New to Fishbowl?
unlock all discussions on Fishbowl.





just refer to machine learning as a trained monkey, works every time
I’d walk the client through the concept of how ensembles can improve predictive power. Start off with something they know (a decision tree) and how maths can be used to make an optimal one. Then move on to the evidence there is on how using many of these decision trees (random forest) on slightly different data (through randomisation) can help cater for the veracity there is in data / models. Possibly then you can elaborate on other successful methods that work similarly (XGB)
Same s..t consultants were explaining to me in 2015 :) .. tell me something new
Microsoft PowerBI is made by Microsoft
Oh I should change that to Accenture now lol, thanks for the reminder
Just say Bayesian a lot
Prior beliefs and updated beliefs
Just wear glasses to the meeting. The internet tells me people will think you’re smarter.
Then have a coworker yell in the background as you slow-mo take them off every time there's an opportunity for a pun
2 words...posed in a way that is half question, half propsal...
ENSEMBLE MODEL?!
Deep reinforcement learning with a great video that shows learning progress. You could even do a hand-calculated example on a whiteboard if you want.
A live demo of DALLE-2 and/or GPT-3 and/or sonantic for context, or something else close to whatever context your audience might best relate to.
Huh. Didn't know Lighthouse went this hard. 👏
I've even done some deep RL research before, but hell if I can remember how any of it works now besides drawing the basic agent-environment loop picture
That’s why no one trusts the big 4 for any tech engagement. They make very little effort on actually getting good talent.
They do but talent does not stay around due to projects being non engaging most of the time .. unless you are part of the lower charging offshore or nearshore team
Hierarchical Linear Modeling is about all I got
Like KPMG2 said, make it Bayesian and you're in business lol
Star and snowflake schemas
Inverse probability treatment weighting applied to propensity scores and then used to run a piecewise cox proportional hazards model with time interactions. 🤓
I prefer logistic hazards for greater flexibility. But I guess some people prefer a rigid Cox
I usually talk about getting from MLOps lvl 1 to MLOps lvl 2.
ML ops is like agile though now. Try harder
Coach
General overview about how a transformer architecture works
I spend half my time trying to make things sound simple and easy.
Coach
Moving from traditional segmentation and modeling to network analysis. Leveraging the direction of movement along the edges to further separate customers into desired behavior and concentrating model development on improvement along the largest pools in the worst directions.
It's the back of the napkin visual and the quick transfer of knowledge of a whole host of concepts that pretty much has them filling in the nodes and doing huge chunks of my work for me... if we have to decend into anything it's usually intro to graph.
Am I wrong in saying you could talk about genetic algorithms in code?
Reward good behavior, punish bad?
Disclaimer - not my field of expertise
The Monte Carlo algorithm for predictive analytics vs the accuracy of a dedicated algorithm