Working at the intersection of production ML infrastructure and high-scale backend engineering
| Domain | Impact |
|---|---|
| 🧠 ML Inference | gRPC layer on TF Serving — ↑ throughput, ↓ latency for nonlinear DL models |
| 🗄️ Feature Platform | Varys Feature Store @ Flipkart — real-time ranking at national scale, multi-DC |
| 🔁 Data Pipelines | Spark + Azkaban → Aerospike; ranking features scaled 8 → 40 |
| 📊 ML Observability | Druid cluster — billions of rows of feature telemetry, DS dashboards |
| 🏋️ Scale Readiness | BBD / Opus NFRs, chaos playbooks — zero P0/P1 on peak load |
| 🔤 NLP Research | Hybrid extractive–abstractive summarization pipeline @ Samsung PRISM |
┌──────────────────────────────────────────────────────────────────────────┐
│ │
│ 🛒 Flipkart SDE-II Aug 2025 → Present │
│ 🛒 Flipkart SDE-I Sep 2023 → Aug 2025 │
│ 📚 Byju's MTS Intern Sep 2022 → Sep 2023 │
│ 🔬 Samsung PRISM Research Intern Oct 2021 → Aug 2022 │
│ │
└──────────────────────────────────────────────────────────────────────────┘


