BI Developer Career Guide: Power BI, Tableau, and Interviews

Business users judge a BI developer on two things: whether the number on the report is right, and whether the page opens before they give up and ask for an Excel file. The data model, the DAX, the refresh schedule and the security rules all serve those two outcomes. The rest of this page covers the work, the tools Indian postings name, and how interviews test both.

What BI developers build and maintain

A BI developer turns warehouse tables into a semantic layer (the tables, relationships, measures and security rules that reports sit on) and then into the reports and dashboards people open every day. Requirements come from finance, sales, supply chain and leadership; data comes from a warehouse, an ERP, or sometimes a folder of Excel files nobody wants to own.

One realistic week:

• Monday's scheduled refresh failed because the source database password was rotated and nobody updated the credentials stored for the gateway. You update them and set up a failure alert so the business hears about problems from you first.
• A sales director says the quarter's revenue on your report differs from finance's figure. The measure ignores credit notes; you correct it and write the definition into the measure's description.
• You build a regional P&L page where each regional head sees only their region, using row-level security driven by a mapping table.
• A page takes 20 seconds to load. Performance Analyzer points to one table visual whose measure iterates a large fact table row by row; you rewrite it.
• You retire four near-identical copies of the sales report that different teams had cloned and edited.

Two roles overlap heavily with this one. A data analyst answers questions and recommends action, and may prototype a dashboard to do it; a BI developer builds the version that hundreds of people rely on every Monday, with security, refresh and performance that hold up. A data engineer builds the pipelines and warehouse tables upstream, and the BI developer starts where those tables end. In MSBI shops and smaller companies the line blurs, and you may write SSIS packages and warehouse SQL as well.

The BI toolset, and what each piece is for

Skill	What you use it for	What panels probe
SQL, often T-SQL	Views and stored procedures that feed models	Joins, window functions, CTEs, and indexing for DirectQuery sources
Power Query (M)	Shaping data before it loads	Query folding, merges, parameters, incremental refresh
Data modelling	Star schemas, relationships, a proper date table	Grain, cardinality, filter direction, and why one wide flat table is slow and fragile
DAX	Measures and time intelligence	Row versus filter context, CALCULATE, context transition, iterators such as SUMX
Tableau	Workbooks, LOD expressions, extracts	FIXED, INCLUDE and EXCLUDE; order of operations; extract versus live connections
SSIS, SSRS, SSAS	ETL packages, paginated reports, tabular and multidimensional models	Incremental loads, package error handling, migrating old cubes and reports
Service administration	Workspaces, gateways, deployment pipelines, row-level security, endorsed shared models	Who can see what, and how a change reaches production

Indian postings are often titled by tool: Power BI Developer, Tableau Developer, Qlik Developer, or MSBI Developer, which means the SQL Server trio of SSIS, SSRS and SSAS. Tableau roles weigh visual design and calculated fields more, with data often prepared upstream or in Tableau Prep; Power BI roles more often include modelling and DAX. Newer Power BI postings often mention Microsoft Fabric as well. Differentiators include DAX Studio and Tabular Editor, calculation groups, source control through Power BI project files, and enough warehousing knowledge to design tables yourself. Know one renaming: what Power BI used to call a dataset is now called a semantic model.

Levels and titles, from report builder to BI architect

0–2 years. You build report pages on an existing model, clean data in Power Query, write straightforward measures and SQL views, and follow the team's design standards. Titles include BI Developer, Associate BI Developer, Power BI Developer and Report Developer.

3–5 years. Now you design models from scratch, write DAX that holds up under any combination of filters, implement row-level security, tune slow reports, run deployments from development to production, and lead requirement sessions with business users. Expect Senior BI Developer or BI Consultant.

6+ years. As BI Lead, BI Architect or BI Manager, you choose platform patterns (import or DirectQuery, shared models or report-specific ones), set governance and workspace strategy, plan capacity and licensing with IT, and run migrations.

Where you work changes the job. Services firms run client reporting projects and frequent migrations from older reporting tools to Power BI, which brings broad exposure but less ownership. Product companies and startups hire fewer dedicated BI developers; analysts or analytics engineers often own reporting there, sometimes on Looker, Metabase or Superset. GCCs build enterprise reporting for global finance, sales and supply chain teams, with strict governance and large shared models.

From screening call to live build: the BI interview rounds

1. Screen: which tools you have used, the size of the models you have built, notice period and expectations.
2. Technical round: SQL, DAX and modelling questions, often with a shared screen.
3. Scenario round: a slow report, a security requirement, a refresh that fails overnight.
4. Live build or take-home: a dataset and 45 to 90 minutes to build a one-page report, or a few days for a fuller one.
5. Portfolio walk-through: a report you built, explained on public or anonymised data.
6. Manager or client round: services firms often add an interview with the client before placing you on their project.

DAX, modelling and performance questions to prepare

1. "When is a calculated column the wrong choice?" Columns are computed at refresh and stored, using memory; measures are evaluated at query time in the current filter context. Use columns to slice and group, measures to aggregate.
2. "Your same-period-last-year measure returns blank on every row. Why?" Usually there is no proper date table: the relationship uses a datetime column that carries times, the dates come from the fact table with gaps, or the table is not marked as a date table.
3. "What happens when CALCULATE runs inside a calculated column?" Context transition turns the current row into a filter. It is powerful, slow on large tables, and surprising when rows are not unique.
4. "Show each product's share of total sales in a matrix. Now a category slicer is added: should the denominator change?" Compare ALL, ALLSELECTED and REMOVEFILTERS, and ask the business which total they mean.
5. "Budget is set by month and region; actuals arrive by day and store. Build one report that compares them." Two fact tables at different grains, shared dimensions at the coarser grain, and measures that hide budget below month level instead of spreading it silently.
6. "A report page takes 25 seconds. Walk me through the diagnosis." Performance Analyzer to find the slow visual, DAX Studio server timings to separate storage-engine from formula-engine time, VertiPaq Analyzer to find high-cardinality columns; then drop unused columns, split datetimes into date and time, remove bidirectional relationships, and reduce visuals per page.
7. "Regional managers see only their region; the national head sees everything." Dynamic row-level security with a mapping table and USERPRINCIPALNAME(), tested with "View as". Add that in Power BI, row-level security restricts viewers but not workspace admins, members or contributors.
8. "In Tableau, a FIXED LOD ignores the region filter you just added. Why?" FIXED expressions are computed before dimension filters; add the filter to context.
9. "Load new and changed rows into a dimension that keeps history, in SSIS." An incremental load into a Type 2 slowly changing dimension; on large dimensions many teams prefer a T-SQL MERGE to the built-in SCD transformation.

Showing BI work on a resume

Reviewers want to know who used your reports, how large and complex the models were, and what you improved: load time, refresh reliability, security, or the number of duplicate reports retired. The bullets below are illustrative, with placeholder numbers.

• Example: "Rebuilt the sales semantic model from a single 120-column flat table into a star schema with a date table and 40 reusable measures; model size fell from 2.1 GB to 600 MB and the main page's load time from 18 seconds to 3."
• Example: "Implemented dynamic row-level security for 250 field users across 14 regions using an HR-maintained mapping table, replacing 14 separately published copies of the same report."
• Example: "Migrated 85 SSRS and Excel reports to Power BI over two quarters, retired 30 duplicates after a usage review, and set up development, test and production deployment pipelines."

Mistakes to avoid: "Created 50+ dashboards" with no users or outcomes; listing visual types such as bar charts, slicers and KPI cards as skills; no mention of modelling or DAX, which makes reviewers assume drag-and-drop work; and screenshots of client reports in a portfolio. Client data is confidential, so rebuild the idea on public data instead.

Breaking into BI and growing beyond it

Common entry routes are MIS and Excel reporting (Power Query already lives inside Excel), data analyst roles, SQL development, and campus or off-campus hiring into the BI practices of services firms. A sensible order of learning:

1. SQL, including window functions.
2. Power BI Desktop with a proper star schema and date table, with auto date/time switched off.
3. DAX evaluation context, until CALCULATE stops surprising you.
4. The Power BI service: workspaces, gateways, scheduled refresh, row-level security, deployment.
5. Performance tooling, then Tableau or the MSBI stack, whichever your target postings name. uHired's BI developer job pages rank the tools that recent postings mention, which settles it for your city.

Portfolio projects that hold up in a walk-through:

• A public sales or operations dataset modelled as a star schema, with fifteen measures including time intelligence, a two-role row-level security demo, and a note on load time before and after tuning. Publish the project files with a README, as the portfolio guide describes.
• The same dashboard built in Power BI and Tableau, with a short write-up comparing CALCULATE with LOD expressions for one business question.
• A paginated invoice or statement report next to an interactive summary of the same data, showing you know when each format fits.

From here, BI developers grow into BI architecture, move upstream into data engineering or analytics engineering, or towards analysis and data science if the questions interest them more than the plumbing. If you plan to move from a services BI practice to a product company or GCC, this guide to switching covers how to present client work.

Where to find BI roles, and a week of practice

Recent listings are grouped by city: Bengaluru, Hyderabad and Pune, plus India as a whole and remote roles. Over the next week:

• Group your shortlist by tool (Power BI, Tableau, Qlik or MSBI) and keep one CV version per group, opening with a project in that tool; How to Read a Software Job Description covers separating must-haves from wish lists.
• Answer the nine questions above aloud with a model open, so you can demonstrate rather than describe.
• Build a one-page report from a raw CSV in 60 minutes, twice, timing yourself.