FP&A Analyst · MBA Finance
Data & BI Analyst · MBA Finance
I run the monthly cycle from close and consolidation through budget vs. actual and the driver-based rolling forecast. I also build the SQL and Power BI layer underneath it, so the pack assembles itself instead of getting rebuilt by hand every month. My models tie out. My variance work tells an allocation problem apart from an overspending one, and the commentary is written for the person who has to act on it.
I build the data layer and I know what the numbers mean, because I closed the books for four years before I started querying them. I type the schema explicitly, deduplicate in a view, keep the raw table intact as an audit trail, and run materiality logic that cuts 4,815 rows down to the eight worth a meeting.
| Measure | Actual | Basis |
|---|---|---|
| Books closed after month-end | 0 days | 36 of 36 months at Wingspan Couture |
| Blended gross margin | 0% | up from 42%, on pricing and markdown |
| Reporting pack turnaround | 0 days | down from 8 days, rebuilt in Power Query |
| Statutory late-filing penalties | ₹0 | 3 years of GST and statutory deadlines |
| Operating expenditure analysed | ₹0 Cr | 36 months · 13 departments · 4 BUs · 4 regions |
About
MBA in Finance with 4+ years across corporate finance and a venture-backed startup. At Wingspan Couture I was the finance function. As sole adviser to the owner-CEO I ran close, consolidation, GST and the monthly management pack, then turned the variances into pricing and markdown decisions that moved blended gross margin from 42% to 48%.
The reporting took eight days and I assembled it by hand every month. I rebuilt it in Power Query and got it down to two. That is the short answer to why I moved toward data. I had already paid for doing it the slow way.
Since March 2026 I have been building the technical half in public: three end-to-end projects covering dataset design, SQL Server ingestion and cleaning, CTE-based analysis and Power BI delivery. Two of the three run on datasets I generated myself, seeded with the faults real ERP extracts carry, including nulls, duplicates, a day/month swap on import, a mid-period account rename, zero-budget spend and negative credit notes. Every repository says so on the first screen. Cleaning broken data is most of the work, so I wanted projects where the data was actually broken.
What I want next is an FP&A or business finance seat where the forecast is driver-based, someone acts on the variance commentary, and nobody rebuilds the pack by hand.
What I want next is an analytics seat close to finance, where knowing what an open period or a credit note does to a headline number counts for more than another dashboard.
Competencies
Three ratings, applied honestly. Owned means years of production responsibility. Applied means I have shipped it end to end in a documented build. Developing means real capability that I picked up recently.
Python is rated Applied rather than Owned because I use it for dataset generation, cleaning and preparation, not for production pipelines. I am working through Power BI financial reporting and DAX in more depth.
Experience
Selected projects
Each project runs from raw data through to a recommendation. Where I generated the dataset instead of sourcing it, the card and the repository both say so.
Python · SQL Server · Excel · Power BI
A business closes 4% over budget on ₹3,758 Cr of operating expenditure. That looks tolerable, so nobody digs further.
Built the pipeline across 36 months, 13 departments and 14 cost categories. The dataset was generated in Python carrying the faults real ERP extracts have, then loaded into SQL Server under an explicitly typed schema and cleaned with ROW_NUMBER deduplication and a DATEFROMPARTS repair of a day/month swap. The raw table stays intact as an audit trail. Nine CTE-based queries feed an Excel variance bridge and a Power BI dashboard whose DAX replicates the SQL open-period exclusion. Without that exclusion the headline reads 0.56% instead of 4.01%.
The 4% was a net figure. ₹239.7 Cr of overspend sat against ₹88.9 Cr of underspend, which puts absolute variance at 8.7% of budget. The business had an allocation problem rather than an overspending one. Two of the largest variances turned out to be single events instead of trends, which only showed up because each cost line was compared to its own monthly average rather than sorted by size.
SQL Server · Excel · Power BI
A multi-channel apparel retailer discounts to move stock and needs to know where discounting stops paying for itself.
Worked roughly 9,600 order lines at transaction level in SQL using LAG, RANK and ROW_NUMBER, standardising category labels, removing duplicates and keeping returns as negative values instead of deleting them. That view feeds six downstream analyses. On top of it sits a two-page Power BI dashboard built on a dedicated date table, with explicit YoY and MoM measures and slicers synced across category, region and channel, plus an Excel layer carrying a FORECAST.ETS seasonal projection and three discount scenarios.
Gross margin holds near 45% at full price and falls to 9.5% once discounting passes 30%, and most of that sits in end-of-season events. Footwear is the second largest category by revenue at ₹41.6 lakh and the lowest margin of the majors at 30.4%. At roughly 28% opex, a three-point margin move swings operating profit by about 45%.
Advanced Excel · Mphasis Ltd.
Historical statements tell you what happened. They do not tell you what happens if growth slows or margins compress.
Rebuilt the FY24 to FY26 consolidated financials as a linked income statement, balance sheet and cash flow, tied to the audited FY26 annual report, with the historical balance check landing at 0.00. The model then extends into an FY27 to FY29 driver-based forecast with a CHOOSE/MATCH Base, Bull and Bear switch that flows revenue growth, margin and working capital assumptions through all three statements at once.
The business consumes cash as it grows. I documented six limitations explicitly, including the fact that forecast cash acts as a balancing plug, which makes the forecast-period balance check an arithmetic identity rather than a validation. The historical tie-out is the real check, and the model says so in writing.
Education & certifications
MACFAST, Mahatma Gandhi University
Mahatma Gandhi University
Udemy
Udemy
Anthropic
Contact
I am open to FP&A, business finance, financial analyst, data analyst and business analyst roles in Bengaluru, remote or on relocation, and I am available for late-evening calls with US and EMEA teams.
mrgokulrnair@gmail.com