Principal Data Scientist

Location: Hydrabad / Bangalore, India

Type: Full-Time | Immediate Joining Preferred

CTC: Competitive

About YAL.ai

YAL.ai (Your Alternative Life) is a next-generation communication and discovery platform that redefines how people connect, interact, and collaborate. We integrate cutting-edge AI into every layer of interaction from multilingual ASR and fraud detection to personalized discovery and recommendation systems. Built on a Zero Trust Architecture, YAL.ai is designed for security, privacy, and intelligence at scale. With our tagline “Where AI Meets Integrity”, we’re creating an ecosystem where discovery is trustworthy, real-time, and hyper-personalized, powering millions of meaningful connections.

Role Overview

We are seeking a Principal Data Scientist (Discovery & Recommendation Systems) to lead the design, research, and scaling of YAL.ai’s core recommendation engine. This is a senior, leadership-level role where you will architect and implement end-to-end discovery pipelines, spanning profile recommendations, semantic matching, community discovery, and real-time personalization. You will be responsible for strategy + execution, driving innovation in graph-based learning, deep retrieval, personalization algorithms, and scalable re-ranking pipelines to deliver state-of-the-art discovery experiences inside YAL.ai’s platform.

Key Responsibilities

- Architect and lead the development of YAL.ai’s recommendation engine for user-user, user-group, and topic discovery.
- Design scalable retrieval + ranking pipelines, integrating dense embeddings, sparse signals, and graph-based features.
- Drive innovation in graph learning (Node2Vec, DeepWalk, GNNs) and semantic similarity (Twin-BERT, SBERT, cross-encoders) for matching and personalization.
- Build cold-start solutions using hybrid approaches (metadata + embeddings + activity signals).
- Develop approximate nearest neighbor (ANN) search at scale (FAISS, HNSW, ScaNN) for sub-100ms retrieval latency.
- Lead re-ranking strategies combining behavioral signals, diversity, trust, and multilingual fairness.
- Integrate multilingual NLP and cross-lingual embeddings to enable code-mixed and Indic language discovery.
- Own offline evaluation metrics (NDCG, Recall@K, MAP) and online A/B testing frameworks for recommendation quality.
- Mentor data scientists, set research direction, and collaborate closely with engineering/product to bring research-grade models into production at scale.

Required Technical Skills

- 8–12 years of experience in applied ML/DS with at least 5+ years in recommendation systems.
- Proven expertise in recommender algorithms (collaborative filtering, content-based, hybrid, contextual bandits, sequence-aware recommenders).
- Hands-on with ANN search libraries (FAISS, HNSW, Milvus, ScaNN) and vector databases.
- Deep experience with transformer-based embeddings for semantic retrieval.
- Strong background in graph-based learning (Node2Vec, DeepWalk, GNNs) for social/user graph discovery.
- Experience with ranking/re-ranking systems (pairwise, listwise ranking, LambdaMART, DLRM, neural rankers).
- Solid foundation in evaluation metrics (NDCG, MRR, AUC, diversity, coverage) and online experimentation (A/B, interleaving).
- Expert in Python, PyTorch/TensorFlow, HuggingFace, and scalable ML pipelines.
- Experience optimizing models for real-time inference (distillation, quantization, batching, async serving).

Qualifications

- Master’s or PhD in Computer Science, AI, Data Science, or related field.
- Strong academic or research background in recommendation systems, ranking models, or graph learning.
- Publications in RecSys, WWW, KDD, SIGIR, ACL, EMNLP, NeurIPS, ICML highly valued.
- Demonstrated ability to bring recommendation models from research to production at scale.

Experience

- 8–12 years total, with significant leadership in recommendation systems.
- Proven track record in real-time, large-scale recommendation pipelines.
- Experience handling multilingual and user-generated data in production.
- Prior work in consumer apps, discovery platforms, or personalization systems strongly preferred.

Bonus If You Have

- Built large-scale social or interest graph recommenders.
- Experience with bandit algorithms or reinforcement learning in recommendations.
- Exposure to trust modeling, bias mitigation, and fairness in recommender systems.
- Hands-on with Indic and code-mixed language recommendation pipelines.
- Contributions to open-source recsys/NLP projects or strong Kaggle performance in recsys/NLP.

How to Apply

- Apply directly via LinkedIn or DM us or
- Send your CV + work samples (GitHub, papers, demos) to hire.ai@yal.chat with Subject: [Principal Data Scientist – Discovery & Recommendation | Your Name]