We're hiring: someone to make our supply chain actually predict things ?

Company: Borosil Limited Team: Sales Transformation & IT – (Supply Chain)

Location: Gurugram (on-site, with external visits)

Level: Data Scientist Exp: 3–4 yrs

Type: Full-time The vibe

Right now, supply chain runs on Data — reports get made, dashboards get refreshed, everyone nods in the review. Cool. We want to skip past that.

We're looking for someone who can take fill rates, freight costs, warehouse spend, and inventory that's been sitting around too long — and actually predict what happens next, before it happens. Not another report nobody opens twice. Real forecasting, real optimization, real ₹ saved.

If "descriptive analytics" makes you yawn and "predictive/prescriptive" makes you lean in — keep reading.

What you'll actually be doing

Performance Metrics & SLA monitoring — but make it predictive

- Own fill rate and TAT tracking across the network, not as a rear-view mirror but as an early-warning system
- Build models that flag an SLA breach before it happens, not in next week's report
- Send scorecards to warehouses that people actually act on

Freight & network analytics

- Deep-dive freight by channel, category, distributor, customer — FTL/PTL splits, route costs, vehicle optimization, the whole map
- Run 4–5 years of freight trend analysis and actually use it
- Apply real optimization/ML (route optimization, load consolidation) to bring cost per unit down, not just report that it went up

Warehouse & cost apportionment

- Cost per sq. ft., cost per box, cost per CFT — know these cold, by channel and category
- Chase down slow-moving and non-moving stock like it owes you money
- ABC/XYZ inventory analysis, pallet occupancy, a discontinued SKU dashboard with aging built in
- Build demand forecasting and inventory optimization models that actually shrink the slow-moving pile

Dashboarding & predictive analytics

- Sales Fill Rate,



Stock Transfer Fill Rate, Inventory Flow, Returns (with root cause, not just totals), Inventory Audit, Customer TAT — dashboards people open because they need to, not because someone told them to
- Layer demand forecasting, anomaly detection, and SLA-risk scoring on top so the MIS stops just reporting and starts warning

How we'll know you're crushing it

What we track

What "good" looks like

? Freight cost/unit

Trending down, with your fingerprints on why

? SLA breach prediction

Catching breaches before they happen, not logging them after

? Slow/non-moving stock

Shrinking, not just monitored

? Dashboard adoption

Warehouse and category teams actually opening them

⏱️ Forecast accuracy

Numbers that hold up when reality shows up

What you need to bring

Education Bachelor's/Master's in Data Science, Statistics, CS, Operations Research, Engineering, or something equally quant-heavy.

Must-haves

- Real stats chops — hypothesis testing, regression, time-series, probability, not just the definitions
- ML that ships: classification, regression, clustering, forecasting (ARIMA, Prophet, XGBoost)
- Demand forecasting and inventory optimization modeling, done for real, not just in a course
- Route/network optimization — linear programming, heuristics
- Anomaly detection for SLA/TAT breaches and cost outliers
- Python or R (pandas, NumPy, scikit-learn, statsmodels) + advanced SQL — joins, CTEs, window functions, on genuinely large transactional data
- Advanced Excel is non-negotiable — Pivot Tables, Power Query, complex formulas; VBA is a bonus




- Comfortable building interactive dashboards, DAX/Power Query for data modeling

Nice-to-haves

- ETL pipelines, data warehousing concepts
- Cloud data platforms — Azure/GCP/AWS
- SAP HANA, WMS, or logistics/ERP exposure
- A/B testing and experimental design for process changes

Vibe check (soft skills)

- You can eyeball a supply chain number and immediately know if it smells wrong
- You're not scared of messy, multi-year transactional data — it will never be clean, make peace with it now
- You can walk into a room with warehouse and category teams and actually get them to listen
- High ownership, low hand-holding

Domain knowledge that'll make this click faster

- FTL/PTL, freight cost structures, warehouse cost apportionment, ABC/XYZ inventory classification, fill rate and TAT metrics — you should already speak this language
- Multi-channel distribution — General Trade, Modern Trade, E-commerce/Q-commerce, D2C — bonus points if you already get why they behave differently

Who tends to crush this role FMCG or recent-age companies — Amazon, Flipkart, Zomato and the like — are a strong background for this. If you've wrangled real operational data at scale and turned it into a model that actually changed a decision, you're exactly who we want.

Why this role doesn't suck

- Your models don't sit in a notebook — they touch warehouse ops, freight spend, and inventory decisions that move real ₹
- You get to build the predictive layer on top of MIS that's existed for years but never looked forward
- Direct line to warehouse and category teams, not five layers of translation
- You're building this from "descriptive" to "predictive/prescriptive" — greenfield, not maintenance mode

If you'd rather predict the breach than write the postmortem, let's talk. ?
- Email your cv to [email protected] or [email protected]

📌 Data Scientist - Supply Chain (Gurugram)
🏢 Borosil
📍 Gurugram