Data Scientist at Myntra (Ads Ranking). M.Tech CSE from IIIT Delhi. I build production ML systems, including feature pipelines, ranking models, and evaluation harnesses.
| Year | Publication & Status | Venue |
|---|---|---|
| 2026 | BlackboxNLP 2026, Co-located with EMNLP 2026 in Budapest, Hungary on October 29th, 2026 | |
| 2024 | 2024 IEEE 9th International Conference for Convergence in Technology (I2CT) |
Myntra · Bengaluru, Karnataka
MIDAS Lab, IIIT Delhi · New Delhi
Classifies depression and anxiety symptoms from internet memes by combining OCR text, Qwen2.5-VL-7B-generated figurative reasoning, and ResNet-50 visual features, fused through a MentalBART encoder with retrieval-augmented context.
Deployed DeathStarBench's 20+ service social network across Docker Compose, k3s, and GKE to benchmark latency and resource utilization under three replica/node topologies. Because Prometheus had nothing to scrape and Jaeger shipped broken, I rebuilt observability on eBPF-based Pixie and repaired Jaeger by hand.
Long-form engineering deep-dives into multimodal AI, edge computer vision, and distributed systems benchmarking.
Field notes, architectural essays, and mental models on production machine learning systems, ranking, and infra.
Video walkthroughs, research paper breakdowns, and hands-on demonstrations of machine learning systems.
He has strong development skills and an impressive ability to understand problem statements clearly, which allows him to deliver effective and well-structured solutions. Punctual, sincere, and highly dedicated, as well as a great team player who collaborates effectively and contributes positively.
Dhruv is one of the most dedicated and hardworking students, consistently outcome-oriented and punctual. He has worked under me for many research projects in AI, ML, LLMs, Computer Vision and Data Science.