Interview Prep · ML Systems Lab

Nykaa Senior Data Scientist / MLE Interview Guide

Nykaa hires senior DS / MLE roles for recommendation, search, personalisation, inventory forecasting, and fraud. The interview tests recommendation depth, fashion/beauty domain understanding, and the ability to drive growth via ranking in a content-rich vertical. This guide covers it.

Nykaa is India's largest beauty and fashion e-commerce player and runs a sophisticated personalisation stack: ranking, search, recommendation, image-based discovery, inventory forecasting. Senior DS / MLE roles split between Nykaa.com (beauty) and Nykaa Fashion, with shared infrastructure.

The loop structure (Senior DS / MLE)

Standard loop: recruiter screen → analytics case + SQL → ML / recommendation depth → ML system design → behavioural.

Round-by-round breakdown

Round 1 — Recruiter screen. 30 min.

Round 2 — Analytics case + SQL. 75 minutes. Case: drive incremental revenue through ranking. SQL on consumer e-commerce schema (sessions, conversions, baskets, returns).

Round 3 — ML / recommendation depth. 75 minutes. Recommender depth: candidate generation vs ranking, multi-task ranking (CTR + AddToCart + Purchase), cold-start strategies, position bias, exploration. Image recommendation: when do you use image embeddings vs collaborative filtering.

Round 4 — ML system design. 60 minutes. Design Nykaa's homepage personalisation, or Nykaa search ranking, or inventory forecasting per SKU per warehouse.

Round 5 — Behavioural. 45 minutes.

What Nykaa weights distinctively

1. Recommender depth. Multi-task ranking is the bread and butter. Position bias correction matters. 2. Domain context. Beauty / fashion has unique signals (skin tone matching, occasion-based purchasing, return patterns). Senior MLEs reason about these. 3. Inventory and supply integration. Recommendation affects inventory dynamics. Out-of-stock items rank lower; this creates feedback effects. 4. Image features. Visual search and image-based recommendation matter more here than in most verticals.

Top 10 questions Nykaa senior DS / MLE candidates face

1. "Design Nykaa's homepage personalisation. Goal: incremental revenue per visit." 2. "Multi-task ranking — CTR, AddToCart, Purchase. How do you train and what loss do you use?" 3. "Cold-start: a new SKU just launched, how does it get ranked?" 4. "Position bias — explain why naive CTR-only training is biased and how you'd fix it." 5. "Image-based recommendation vs collaborative filtering — when does each win for beauty?" 6. "Walk through diagnosis of falling AddToCart rates on a featured category page." 7. "Inventory feedback loop: recommendations affect inventory, inventory affects recommendations. How do you design for stability?" 8. "Design search ranking for Nykaa: typo correction, semantic matching, freshness, price." 9. "Returns prediction — when do you flag a high-return-risk product to a customer?" 10. "A/B test on ranking showed +5% CTR but -2% revenue. What happened?"

The prep path through MSL

  • Tier 5 (Evaluation) — calibration, position bias.
  • Tier 7 (Production Engineering) — all 5 posts.
  • Tier 8 (Monitoring) — all 5 posts.
  • Posts 48 (Recommender Feedback Loops), 70-72 (RecSys series), 96 (Bandits for exploration).
  • Tier 9 (DS & Causal) — Posts 91 (Network Effects), 93 (Metrics).
  • Practice: MockInterview with Nykaa JD pasted.
  • Common failure modes

  • Surface recommender answers without depth on multi-task, position bias, feedback loops.
  • Ignoring inventory effects on ranking.
  • Inability to articulate revenue vs CTR conflicts.
  • Compensation

    Nykaa Senior DS / MLE in 2026 ranges roughly ₹30 lakh – ₹55 lakh for 5-7 YOE. Lead roles ₹70 lakh+. Below adtech / fintech rates but with strong consumer ML depth and clear growth path.

    Continue interactively
    Read this post inside ML Systems Lab — with Simplify toggle, interview Q&As, inline glossary, and the MLE Path forward pointer.
    Open in MSL →