I am a seasoned Data Scientist/Analyst with a rich experience spanning two years in the field. Over the past year, I have proudly contributed my skills and expertise to the dynamic teams of both Data Science (Personalisation - ML) and Data Engineering at McAfee. My journey has been marked by a relentless pursuit of excellence and a passion for transforming data into meaningful, concrete insights.

My proficiency in SQL, Python, Spark, my ability to figure out a way around blockers on the go, and my decisiveness with sound reasoning has been the cornerstone of my success, enabling me to extract, analyze, and interpret data with precision. My adeptness in these tools has empowered me to spearhead feature engineering projects, playing a pivotal role in enhancing the value of data and optimizing its use for strategic decision-making.

One of my significant accomplishments includes being a key contributor to the establishment of a feature store at McAfee. 
As we all know, from start to end, 70% of time in most Machine Learning projects is spent on finding out correct POCs within the company, communicating with them to acquiring the correct data and also understanding it, predicting probable metrics to be obtained from it, and then performing the pre-processing. 
Developing a feature store involved a lot of Feature Engineering, gaining close familiarity with the raw data, understanding the business and performing a lot of SQL intricacies, not to forget a lot of internal communication. The objective was to cut down on the preprocessing time and provide Data Scientists 
with data they could rapidly build prototypes upon without having to worry about the entire logistics of how it came to be. The nature of this 'pre-processed' data was that it could support a lot of Data Science efforts, since this data captured the essential behavior of our customers. 

This innovative solution has not only streamlined workflows but has also facilitated the acceleration of data-driven innovation within the organization.

Motivated by my unwavering commitment to continuous growth and a thirst for knowledge, I am eager to expand my horizons into the realms of Data Engineering and Data Science. Drawing upon my foundation in analytics and my proficiency in key technologies, I am excited to take on new challenges and leverage my skill set to drive transformative insights and solutions.

In essence, my journey reflects a fusion of experience, technical prowess, and an unrelenting desire to create value from data.

PhonePe — Decision Scientist (Round 1) | Interview Experience

SQL fundamentals + product sense in 30 minutes.

━━━━━━━━━━━━━━━━━━━━━━━━━━

𝗦𝗤𝗟 𝗤𝘂𝗲𝘀𝘁𝗶𝗼𝗻𝘀:

𝟭. Running sum by city
SUM(month_total) OVER(
PARTITION BY city
ORDER BY month
)

𝟮. INNER vs LEFT JOIN with duplicates
Tables: 111223445 and 111233455
• INNER JOIN output: 1112345
• LEFT JOIN output: 111223445

Testing: Do you understand join behavior?

𝟯. Anti-join (rows in T1 not in T2)
SELECT t1.*
FROM t1
LEFT JOIN t2 ON t1.col = t2.col
WHERE t2.col IS NULL

━━━━━━━━━━━━━━━━━━━━━━━━━━

𝗣𝗿𝗼𝗱𝘂𝗰𝘁 𝗖𝗮𝘀𝗲:
"Cart-to-order conversion dropped 40% on Swiggy. How do you localize?"

My approach:

𝟭. Check the funnel
Cart → Checkout → Payment → Order

𝟮. Key metrics
Cart abandonment, payment failures, error rates, recent UI changes

𝟯. Hypothesize causes
Payment gateway issues, app crashes, promo failures, price sensitivity

𝟰. Segment to localize
Geography, customer type (new vs repeat), time of day, device, payment method

Use: A/B test data, heatmaps, session recordings

━━━━━━━━━━━━━━━━━━━━━━━━━━

They're not testing for the "right" answer.

They're testing systematic debugging.

Know your SQL fundamentals. Think in funnels. Debug like a detective.

P.S. Join 500+ data professionals sharing interview experiences, referrals, job opportunities, and resources: https://lnkd.in/gp7u5VKY

#DataScience #ProductAnalytics #InterviewExperience #PhonePe #SQL

Sidharth Kriplani

Senior Data Analyst at Uber (through Indium)

Meesho — Senior Business Analyst (Tech Round 2) | Interview Experience
Two SQL questions. Both practical.

━━━━━━━━━━━━━━━━━━━━━━━━━━

𝗤𝘂𝗲𝘀𝘁𝗶𝗼𝗻 𝟭: 𝗦𝘄𝗶𝗴𝗴𝘆 𝗢𝗿𝗱𝗲𝗿𝘀 𝗔𝗻𝗮𝗹𝘆𝘀𝗶𝘀
"Design a schema for orders, then calculate the daily percentage share of orders from 'good users'."
Good user = >10 orders in last 2 weeks

𝗔𝗽𝗽𝗿𝗼𝗮𝗰𝗵:
1. Design orders table
   • order_id, user_id, order_date, order_value, store_id

2. Identify good users
   • COUNT orders per user in rolling 14-day window
   • Flag users with >10 orders

3. Calculate daily share
   • Daily total orders
   • Daily orders from good users
   • percentage = (good_user_orders / total_orders) * 100

Key: Window function for rolling 14-day count + daily aggregation

━━━━━━━━━━━━━━━━━━━━━━━━━━

𝗤𝘂𝗲𝘀𝘁𝗶𝗼𝗻 𝟮: 𝗚𝗿𝗼𝗰𝗲𝗿𝘆 𝗦𝘁𝗼𝗿𝗲 𝗙𝗼𝗼𝘁𝗳𝗮𝗹𝗹
Given: cust_id, timestamp, remark ('in' or 'out')
"Calculate average time spent per customer."

𝗔𝗽𝗽𝗿𝗼𝗮𝗰𝗵:
1. Self-join on cust_id
   • Match 'in' with next 'out' for same customer

2. Calculate time difference
   • TIMESTAMPDIFF(timestamp_out - timestamp_in)

3. Average per customer
   • GROUP BY cust_id, AVG(time_spent)

Edge case: Handle incomplete visits (no matching 'out')

━━━━━━━━━━━━━━━━━━━━━━━━━━

Both questions tested:
✓ Schema design thinking
✓ Window functions
✓ Self-joins
✓ Time-based calculations
✓ Handling edge cases

Senior BA roles aren't just about reports.

You need to design data models and write production-quality SQL.

P.S. Join 500+ data professionals sharing interview experiences, referrals, job opportunities, and resources: https://lnkd.in/gp7u5VKY

#DataAnalytics #SQL #InterviewExperience #Meesho #BusinessAnalyst

I went from beginner to advanced SQL in one week.

Same pressure at my current job.
SQL? Easy to ramp up fast.
Python? Here's what you actually need to hit intermediate level:

━━━━━━━━━━━━━━━━━━━━━━━━━━

𝗧𝗵𝗲 𝟭𝟴 𝗰𝗼𝗻𝗰𝗲𝗽𝘁𝘀:

𝗗𝗮𝘁𝗮 𝗦𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲𝘀:
1. Lists
2. Tuples
3. Dictionaries
4. Sets
5. Strings

𝗖𝗼𝗺𝗽𝗿𝗲𝗵𝗲𝗻𝘀𝗶𝗼𝗻𝘀:
6. List Comprehensions
7. Dictionary Comprehensions

𝗙𝘂𝗻𝗰𝘁𝗶𝗼𝗻𝘀:
8. Functions
9. Lambda Functions
10. Recursive Functions

𝗖𝗼𝗻𝘁𝗿𝗼𝗹 𝗙𝗹𝗼𝘄:
11. If-Else Statements
12. For Loops
13. While Loops

𝗕𝘂𝗶𝗹𝘁-𝗶𝗻 𝗙𝘂𝗻𝗰𝘁𝗶𝗼𝗻𝘀:
14. Enumerate
15. Zip
16. Map
17. Filter
18. Reduce

━━━━━━━━━━━━━━━━━━━━━━━━━━

That's it.

Master these 18, and you're at intermediate Python.

Not "learn Python in 30 days."
Not "100 concepts to know."

Just 18 concepts that show up in 90% of data analytics work.

━━━━━━━━━━━━━━━━━━━━━━━━━━

Bookmark this. Go through each one. Make sure you can use them without Googling.

That's interview-ready Python.

P.S. Join 500+ data professionals sharing interview experiences, referrals, job opportunities, and resources: https://lnkd.in/gp7u5VKY

#Python #DataAnalytics #InterviewPrep #DataScience #Coding

You're probably leaving 30-40% on the table in salary negotiations.
Not because you lack skills.
Because you lack data.

Here's the ChatGPT prompt I use to research compensation before every application:

━━━━━━━━━━━━━━━━━━━━━━━━━━━

𝗪𝗵𝗮𝘁 𝗶𝘁 𝗱𝗼𝗲𝘀:

Give it:
• Company name + Role + JD
• Your current CTC

Get back:
✓ Base pay estimate (min → max)
✓ Total comp breakdown
✓ Confidence score (how reliable the data is)
✓ Evidence with citations
✓ Recommended ""Ask Band"" for negotiation

It researches Levels.fyi, Glassdoor, Blind, LinkedIn Salary, and market comps automatically.

━━━━━━━━━━━━━━━━━━━━━━━━━━━

𝗥𝗲𝗮𝗹 𝗲𝘅𝗮𝗺𝗽𝗹𝗲:
Company offered: 18L
Research showed: 22-28L
My ask: 25L
Final: 24L + signing bonus
Without data? I would've accepted 18L.

━━━━━━━━━━━━━━━━━━━━━━━━━━━

Before I apply to any job, I run this prompt.
Before any negotiation, I have the numbers ready.

Most candidates guess their worth.
You'll know yours.

Here is the full prompt: 
https://lnkd.in/ghEyg8s4

Copy it. Use it. Get paid what you're worth.

P.S. Join 500+ data professionals networking and sharing resources: https://lnkd.in/gp7u5VKY

#JobSearch #SalaryNegotiation #CareerGrowth #DataJobs #Compensation

I never sat for college placements.
Figured out job hunting the hard way.

Here's what actually works:

━━━━━━━━━━━━━━━━━━━━━━━━━━

Most people click "Easy Apply" and wait.
I did that once. Got nowhere.

Then I realized: LinkedIn isn't a job board.
It's a networking platform.

━━━━━━━━━━━━━━━━━━━━━━━━━━

𝗠𝘆 𝟯-𝘀𝘁𝗲𝗽 𝗿𝗲𝗳𝗲𝗿𝗿𝗮𝗹 𝘀𝘆𝘀𝘁𝗲𝗺:

𝟭. 𝗙𝗶𝗻𝗱 𝘁𝗵𝗲 𝗿𝗶𝗴𝗵𝘁 𝗷𝗼𝗯
Google, LinkedIn, Naukri, company career pages.
When you find a good match → don't apply yet.

𝟮. 𝗙𝗶𝗻𝗱 𝘁𝗵𝗲 𝗿𝗶𝗴𝗵𝘁 𝗽𝗲𝗼𝗽𝗹𝗲
Go to the company's LinkedIn page → Click "Employees"
Look for:
• People with the same role as the job posting
• Leadership in that vertical

Not recruiters. Not random employees.
People who actually know the team.

𝟯. 𝗔𝘀𝗸 𝗱𝗶𝗿𝗲𝗰𝘁𝗹𝘆
No small talk. Send a connection request, wait for them to accept it, and as soon as you get the notification send them something like this:

"Hi [Name],

Hope all is well with you! I came across the [Role] role at [Company]. Would you be open to submitting a referral for me?

Happy to chat more if you have the time as well. Looking forward to hearing from you.

Thanks,
[Your Name]"

━━━━━━━━━━━━━━━━━━━━━━━━━━

This is how I got every job I've had.

Not through "Easy Apply."
Through referrals.

Most people overthink it.
Find the job. Find the person. Ask directly.
That's it.

P.S. Join 500+ data professionals sharing interview experiences, referrals, job opportunities, and resources: https://lnkd.in/gp7u5VKY

#JobSearch #Referrals #CareerGrowth #LinkedIn #JobHunting

You're wasting time on jobs you'll never get.

Or worse — applying to jobs beneath your level.

Here's the ChatGPT prompt I use to score every job automatically:

━━━━━━━━━━━━━━━━━━━━━━━━━━━

𝗪𝗵𝗮𝘁 𝗶𝘁 𝗱𝗼𝗲𝘀:
Give it:
• Your resume
• Any job description

Get back in ONE line:
✓ Fit score (0-10)
✓ Action tag (what to do)

Example outputs:
9.2/10 – referral_only_no_apply
7.8/10 – apply_tailored
4.3/10 – IGNORE

━━━━━━━━━━━━━━━━━━━━━━━━━━━

𝗧𝗵𝗲 𝗮𝗰𝘁𝗶𝗼𝗻 𝘁𝗮𝗴𝘀:
• 9.0-10: referral_only (perfect match, get a referral)
• 8.0-8.9: referral_attempt (try for referral, then apply)
• 7.0-7.9: apply_tailored (good fit, tailor resume)
• 5.0-6.9: cold_apply (okay fit, quick apply)
• 0.0-4.9: IGNORE (skip it)

━━━━━━━━━━━━━━━━━━━━━━━━━━━

Before this:
• Read 50 JDs, guess which to apply to
• Wasted referrals on mediocre matches
• Applied to roles I had no shot at

After:
• 30 seconds to score any job
• Only ask referrals for 8+ scores
• Skip mismatches immediately

No more decision fatigue.
No more wasted referrals.

━━━━━━━━━━━━━━━━━━━━━━━━━━━

Here is the prompt:
https://lnkd.in/gJ9dWuab

Paste your resume once, then score every job in seconds.

P.S. Join 500+ data professionals networking and sharing resources: https://lnkd.in/gp7u5VKY

#JobSearch #CareerGrowth #DataJobs #JobHunting #Automation

taplio

Sidharth Kriplani's Linkedin Analytics