Analytics Engineer · open to product-company roles

The pipeline between data and decisions.

I'm Rishit Sitholey. I design ETL/ELT pipelines, Snowflake warehouses, and the dimensional models behind Power BI dashboards that leadership actually trusts.

View case studies ↓ Download résumé
SQLPythonSnowflakePower BIAWS S3
01 · About

A product profile,
not a bio.

Portrait of Rishit Sitholey
Lucknow, India · IST

Two years ago I started as a data analyst cleaning lead lists. Today I own the full path at Appex Multisol — from raw S3 drops and operational databases, through Snowflake staging and star-schema modeling, to the Power BI dashboards teams open every Monday morning.

The through-line: I treat data like a product. Declared grains, validated loads, documented models, metrics defined once. If a number reaches a stakeholder, I can trace exactly where it came from.

0
years in data
0
records · largest dataset
0
dashboards shipped
0
source tables per load
B.Tech ECE · ABES Engineering College Analyst → Analytics Engineer · promoted Jan 2026 Learning: Databricks · PySpark · Tableau
02 · Experience

Milestones,
in sequence.

AM

Analytics Engineer

Jan 2026 — Present
Appex Multisol Pvt Ltd · Lucknow ● current
  • Built ETL/ELT pipelines in SQL and Python loading 15–25 source tables into a Snowflake data warehouse for analytics-ready reporting.
  • Designed star-schema dimensional models in Snowflake powering 5–8 Power BI dashboards, improving how teams access core metrics.
  • Ran scheduled and on-demand jobs processing 100K–500K records, with validation checks and monitoring to catch quality issues before dashboards.
AM

Data Analyst

Jan 2025 — Dec 2025
Appex Multisol Pvt Ltd · Lucknow
  • Processed datasets of 100K+ records, structuring staging and warehouse tables for reliable, repeatable reporting.
  • Extracted and loaded data from multiple sources including AWS S3, transforming it into fact and dimension tables.
  • Partnered with stakeholders to translate reporting requirements into the datasets and models that answer them.
BJ

Data Analyst

Jun 2024 — Dec 2024
Bravoji.com · Remote
  • Built 10+ Power BI dashboards (DAX) tracking KPIs, ROI, and campaigns — contributing to a reported 15% revenue increase.
  • Analyzed sales data across 50+ TikTok e-commerce stores to surface trends in revenue, product performance, and campaigns.
  • Cleaned and transformed 50,000+ leads using Power Query to make the data reliable enough for analysis.
PS

Application Support Intern

Apr 2024 — May 2024
PeopleStrong · Gurugram
  • Managed and validated HCM platform data for 500+ employees, maintaining accuracy and system reliability.
  • Resolved 200+ data-related issues by tracing root causes and applying structured fixes.
03 · Projects

Case studies,
not screenshots.

The problem

Analytical questions kept requiring fresh ad-hoc queries against raw, inconsistent extracts — slow to produce, impossible to reproduce. The fix wasn't a better query; it was a warehouse.

The build

Raw data lands untouched, loads into typed staging via COPY INTO, transforms once inside Snowflake — window-function dedupe, conformed types, MERGE-based incremental upserts — and publishes a documented star schema (1 fact + dimensions) as the single interface for analytics.

0records loaded
1+Nfact + dims
4warehouse layers
transform/dedupe.sql
WITH ranked AS (
  SELECT s.*,
    ROW_NUMBER() OVER (
      PARTITION BY order_id
      ORDER BY updated_at DESC) rn
  FROM stg_orders s)
SELECT * FROM ranked WHERE rn = 1;
Impact: repeatable answers (same question, same number, any day), self-serve analytics off a documented model, and validation gates that stop bad data in staging — dashboards only refresh from clean tables.

The problem

Finance, sales, marketing, and supply chain looked at the same transactions through different spreadsheets — and reached different numbers.

The build

2.2M+ records shaped into a dimensional model, imported into Power BI, and every KPI defined once as a central DAX measure. Ten-plus dashboard pages share the model — drill-through connects executive summary to transaction line.

Before

Four teams · four spreadsheets · four versions of "revenue"

→
After

One model · one measure library · one version of the truth

model/measures.dax
Total Revenue = SUM(fct_sales[revenue])

Revenue YoY % =
  DIVIDE([Total Revenue] - [Revenue LY],
         [Revenue LY])

Revenue YTD =
  TOTALYTD([Total Revenue], dim_date[date])
0records
0dashboards
4domains
Impact: when finance asks why their number differs from sales — it can't. Both pages call the same measure.

The problem

Campaign spend across 50+ TikTok stores was steered by gut feel — no unified view of which products, stores, or creatives actually earned budget.

The build

50,000+ leads cleaned and standardized with Power Query, then 10+ DAX-driven dashboards tracking KPIs, ROI, and campaign trends across every store — the reporting the team used to rank spend decisions.

0stores analyzed
0leads cleaned
0reported revenue lift
Impact: gut-feel budget calls became ranked, evidenced ones — the same cleaning → modeling → metrics pattern I've reused ever since.
04 · Skills

A constellation,
not a list.

hover a node · outer dashed ring = currently learning
05 · Contact

Open a channel.
I reply fast.

rishit@contact: ~
or directly: rishitsitholey002@gmail.com · LinkedIn · GitHub