mirror of
https://github.com/dholerobin/Lecture_Notes.git
synced 2025-09-13 13:52:12 +00:00
Add all my notes
This commit is contained in:
213
imperfect_notes/pragy/intro week - Dare to Dream.md
Normal file
213
imperfect_notes/pragy/intro week - Dare to Dream.md
Normal file
@@ -0,0 +1,213 @@
|
||||
Intro Week - 1 - Dare to Dream
|
||||
------------------------------
|
||||
|
||||
- mute all
|
||||
- disable mentee sharing
|
||||
- stop mentee videos
|
||||
- disable screen annotations
|
||||
|
||||
-- --
|
||||
|
||||
Intro
|
||||
-----
|
||||
|
||||
- instructor introduction
|
||||
- myself: pragy, senior soft engg at IB. worked at directi, gateoverflow. IIT-B masters, machine learning. Passion for teaching
|
||||
- anshuman:
|
||||
- facebook messaging - 20M messages per minute
|
||||
- facebook london office - first 100 engineers
|
||||
- interviewbit & scaler
|
||||
- stark diff b/w people hired in US vs India
|
||||
- indian people are theoritical
|
||||
- if you break down and define the task, they will be able to do
|
||||
- as you grow, your job is more of taking vaguely defined problems and solving them
|
||||
- that practical part was missing in indian devs
|
||||
- overview of curriculum
|
||||
- some preview material from different topics
|
||||
- this is a meeting and not a view because interactive
|
||||
- superx content is slightly harder and moves at faster pace. might move faster.
|
||||
- elitex is more indepth with more basics covered
|
||||
- content and curriculum remains exactly the same
|
||||
- can move b/w batches
|
||||
- if struggle with elitex, don't move to superx
|
||||
- backend vs frontend batch - your choice
|
||||
- build things from scrath. detailed curriculum will be shared
|
||||
- classes are recorded. Lecture notes are shared
|
||||
- if you have something urgent, we can share previous batch videos, but don't overstrain yourself. Take things at the correct pace
|
||||
- watching recorded is counted in attendance
|
||||
|
||||
|
||||
-- --
|
||||
|
||||
Presentation
|
||||
------------
|
||||
|
||||
- left guy is fastest man in India - Abhilal
|
||||
- right guy is fastest man in Asian - ...
|
||||
- Usian Bolt - fastest in the world
|
||||
- bolt is more consistent
|
||||
- if you're consistent, you'll get the best results
|
||||
- student example from previous batch
|
||||
- extremely consistent
|
||||
- never missed a class
|
||||
- offer from HCL - 3.5 lakhs
|
||||
- consistent for 1 year. > 20 LPA salary
|
||||
- Google - Naman Bhalla, 3rd tier college. Extremely consistent.
|
||||
- no internship
|
||||
- we first got him an internship then job - 4 months
|
||||
- talent is not why people win
|
||||
- it is hard work and consistency
|
||||
- 1% better everyday is all you need to focus on
|
||||
- we've figured out the long terms goals. we've a pipeline for you. You can focus on the daily tasks, with the assurance that if you follow them, you will end up being an amazing engineer
|
||||
|
||||
- 6 month course
|
||||
- but we will help you forever
|
||||
- scaler talks will continue
|
||||
- always reach out to us for help
|
||||
- reach out to us for referrals
|
||||
|
||||
- will take questions after the lecture
|
||||
|
||||
-- --
|
||||
|
||||
Company Questions
|
||||
-----------------
|
||||
|
||||
- random DS/Algo, design questions that have been asked at companies
|
||||
- not learning, for self evalution
|
||||
- see where we currently stand
|
||||
- we will do another such session 4/4.5 months down the line and see what progress we've made
|
||||
- are you actually spending time in the right direction
|
||||
|
||||
- real world problems - so actual company questions
|
||||
|
||||
-- --
|
||||
|
||||
Facebook Graph
|
||||
--------------
|
||||
|
||||
> friend = edge
|
||||
> represented as a graph
|
||||
> given 2 nodes, find out the degree of separation
|
||||
> unweighted
|
||||
|
||||
2 billions nodes and 2 trillion edges
|
||||
user can have max 5k edges
|
||||
avg has 1k edges
|
||||
|
||||
degree of separation was <= 5, as long as the person had more than 25 friends (not exhaustive, estimate)
|
||||
|
||||
consider everyone has DOS <= 5
|
||||
|
||||
**Sortest path algo**
|
||||
|
||||
will give the correct answer, but nodes is very large
|
||||
|
||||
Let's say BFS - O(V+E) time
|
||||
10^9 computations per second = 33 mins per query
|
||||
|
||||
**Don't go beyond the length of 5**
|
||||
|
||||
but you will still go over everyone with >=25 friends
|
||||
|
||||
|
||||
**Solution**
|
||||
Meet in the middle
|
||||
|
||||
10^6, else ans=5
|
||||
|
||||

|
||||
|
||||
how to check if connected? BBST? Bloom Filter?
|
||||
can use DSU to check if connected for fast return in O(n)
|
||||
or can simply use intersection of lists using hashmap in O(n)
|
||||
or can use sorting in O(n log n)
|
||||
-- --
|
||||
|
||||
|
||||
Google - Running Median
|
||||
-----------------------
|
||||
|
||||
> Given a stream of integers, find the median at every step
|
||||
> more difficult: distributed streams. Have multiple servers and partial data is coming at different servers
|
||||
> cannot pull data into one machine because of memory constraints
|
||||
>
|
||||
|
||||
Median = mid (odd) or avg of 2 mids (even)
|
||||
|
||||
1. sort array and mid number is median - BBST
|
||||
2. store equal number of < and > numbers - 2 heaps
|
||||
|
||||
|
||||
-- --
|
||||
|
||||
Google - Distributed Running Median
|
||||
------------------------------------
|
||||
|
||||
1. ~~median of medians~~. find median of each server. sort the values and find overall median
|
||||
- incorrect
|
||||
- 1 2 (4) 8 10
|
||||
- 1 2 (100) 101 102
|
||||
- median of 4, 100 is 52.
|
||||
- 1 1 2 2 4 8 10 | 100 101 102
|
||||
- 52 is not the median
|
||||
3. each server has BBST
|
||||
4. guess x -> goto each server and count how many numbers are smaller than x and larger than x
|
||||
5. binary search for the answer x
|
||||
|
||||
|
||||
-- --
|
||||
|
||||
DS/Algo 101 - Arrays and Maths
|
||||
HLD 101
|
||||
LLD 101 - OOPs design
|
||||
- give a problem. ask to code.
|
||||
- change requirements. measure the number of changes
|
||||
|
||||
Hacking your interview
|
||||
Celebrity Lecture - CEO, codechef, github, ..
|
||||
|
||||
-- --
|
||||
|
||||
from next week, class on alternate days
|
||||
TAs while solving assignments and homeworks
|
||||
TA sessions
|
||||
(choice) super elite split
|
||||
Mentor sessions (university seniors) + mock interviews every alternate
|
||||
DS Algo 1.5
|
||||
CS basics 1 week
|
||||
HLD - 15 days
|
||||
LLD - 7 days
|
||||
Machine Coding - 8 days
|
||||
Basic of JS, HTML, CSS - 1 week
|
||||
(choice) Full stack vs Backend - 1 month
|
||||
|
||||
-- --
|
||||
|
||||
Placement
|
||||
---------
|
||||
|
||||
- april 196 people - 130+ have min ctc+ have jobs, some are yet to graduate. 18.2 LPA. 52 amazon, 3 google, adobe, 12 microsoft, atlassian, cisco
|
||||
|
||||
|
||||
Referral
|
||||
--------
|
||||
|
||||
choice to stop or continue.
|
||||
|
||||
|
||||
-- --
|
||||
|
||||
How to succeed at Scalar Academy
|
||||
--------------------------------
|
||||
|
||||
- take notes - your own notes
|
||||
- aha
|
||||
- important stuff
|
||||
- theory & formulae
|
||||
- analogies & intuition
|
||||
- discuss with peers
|
||||
- 3 hours everyday
|
||||
- consistency - maintain your streak, even if you're not able to do the stuff completely.
|
||||
- feedback forms (because we use them extensively)
|
||||
- don't have a huge backlog (don't join if major life event)
|
Reference in New Issue
Block a user