Data Scientist - Face To Face - Comcast India LLP

About the Role:

We are seeking an experienced Data Scientist to join our growing team. You will play a key role in building intelligent systems that help reduce alert noise, detect anomalies, correlate events, and proactively surface operational insights across our large-scale streaming infrastructure.

Youll work at the intersection of machine learning, observability, and IT operations, collaborating closely with Platform Engineers, SREs, Incident Managers, Operators and Developers to integrate smart detection and decision logic directly into our operational workflows.

This role offers a unique opportunity to push the boundaries of AI/ML in large-scale operations. We welcome curious minds who want to stay ahead of the curve, bring innovative ideas to life, and improve the reliability of streaming infrastructure that powers millions of users globally.

What Youll Do:

Design and tune machine learning models for event correlation, anomaly detection, alert scoring, and root cause inference

Engineer features to enrich alerts using service relationships, business context, change history, and topological data

Apply NLP and ML techniques to classify and structure logs and unstructured alert messages

Develop and maintain real-time and batch data pipelines to process alerts, metrics, traces, and logs

Use Python, SQL, and time-series query languages (e.g., PromQL) to manipulate and analyze operational data

Collaborate with engineering teams to deploy models via API integrations, automate workflows, and ensure production readiness

Contribute to the development of self-healing automation, diagnostics, and ML-powered decision triggers

Design and validate entropy-based prioritization models to reduce alert fatigue and elevate critical signals

Conduct A/B testing, offline validation, and live performance monitoring of ML models

Build and share clear dashboards, visualizations, and reporting views to support SREs, engineers, and leadership

Participate in incident postmortems, providing ML-driven insights and recommendations for platform improvements

Collaborate on the design of hybrid ML + rule-based systems to support dynamic correlation and intelligent alert grouping

Lead and support innovation efforts including POCs, POVs, and exploration of emerging AI/ML tools and strategies

Demonstrate a proactive, solution-oriented mindset with the ability to navigate ambiguity and learn quickly

Participate in on-call rotations and provide operational support as needed

Qualifications:

Bachelor's or Master's degree in Computer Science, Data Science, Machine Learning, Statistics or a related field

5+ years of experience building and deploying ML solutions in production environments

2+ years working with AIOps, observability, or real-time operations data

Strong coding skills in Python (including pandas, NumPy, Scikit-learn, PyTorch, or TensorFlow)

Experience working with SQL, time-series query languages (e.g., PromQL), and data transformation in pandas or Spark

Familiarity with LLMs, prompt engineering fundamentals, or embedding-based retrieval (e.g., sentence-transformers, vector DBs)

Strong grasp of modern ML techniques including gradient boosting (XGBoost/LightGBM), autoencoders, clustering (e.g., HDBSCAN), and anomaly detection

Experience managing structured + unstructured data, and building features from logs, alerts, metrics, and traces

Familiarity with real-time event processing using tools like Kafka, Kinesis, or Flink

Strong understanding of model evaluation techniques including precision/recall trade-offs, ROC, AUC, calibration

Comfortable working with relational (PostgreSQL), NoSQL (MongoDB), and time-series (InfluxDB, Prometheus) databases

Ability to collaborate effectively with SREs, platform teams, and participate in Agile/DevOps workflows

Clear written and verbal communication skills to present findings to technical and non-technical stakeholders

Comfortable working across Git, Confluence, JIRA, & collaborative agile environments

Nice to Have:

Experience building or contributing to the AIOps platform (e.g., Moogsoft, BigPanda, Datadog, Aisera, Dynatrace, BMC etc.)

Experience working in streaming media, OTT platforms, or large-scale consumer services

Exposure to Infrastructure as Code (Terraform, Pulumi) and modern cloud-native tooling

Working experience with Conviva, Touchstream, Harmonic, New Relic, Prometheus, & event-based alerting tools

Hands-on experience with LLMs in operational contexts (e.g., classification of alert text, log summarization, retrieval-augmented generation)

Familiarity with vector databases (e.g., FAISS, Pinecone, Weaviate) and embeddings-based search for observability data

Experience using MLflow, SageMaker, or Airflow for ML workflow orchestration

Knowledge of LangChain, Haystack, RAG pipelines, or prompt templating libraries

Exposure to MLOps practices (e.g., model monitoring, drift detection, explainability tools like SHAP or LIME)

Experience with containerized model deployment using Docker or Kubernetes

Use of JAX, Hugging Face Transformers, or LLaMA/Claude/Command-R models in experimentation

Experience designing APIs in Python or Go to expose models as services

Cloud proficiency in AWS/GCP, especially for distributed training, storage, or batch inferencing

Contributions to open-source ML or DevOps communities, or participation in AIOps research/benchmarking efforts

Certifications in cloud architecture, ML engineering, or data science specializations