Job Overview
Location
Bengaluru
Job Type
Full-time
Category
Software Engineering
Date Posted
October 27, 2025
Full Job Description
đź“‹ Description
- • Architect and own the end-to-end ML lifecycle that powers the personalized home page for 750 million weekly viewers, ensuring every tile, trailer, and thumbnail is individually ranked in <100 ms.
- • Design ultra-low-latency, high-throughput serving systems that handle millions of concurrent recommendation requests, leveraging GPU inference, distributed caching (Redis, Memcached), and edge nodes to maintain sub-second response times across mobile, web, and smart-TV apps.
- • Build and scale real-time feature pipelines using Spark Structured Streaming and Kafka to compute user embeddings, content embeddings, and contextual signals within seconds of user interaction, enabling dynamic re-ranking as tastes evolve.
- • Lead the creation of an experimentation platform that runs 100+ concurrent A/B tests on ranking algorithms, UI layouts, and notification strategies, automating statistical power calculations, traffic splitting, and guard-rail metric monitoring.
- • Establish rigorous MLOps practices with MLflow for experiment tracking, model registry, CI/CD, and automated rollback; integrate drift, bias, and performance monitors that trigger retraining jobs and alert on-call engineers before viewers notice degradation.
- • Optimize distributed training workflows for deep-learning models (DSSM, Transformers, Two-Tower) on petabyte-scale datasets, employing parameter servers, gradient compression, and mixed-precision training to cut GPU hours by 40 %.
- • Collaborate with Product, Data Science, Design, and Content Ops to translate business goals (watch-time, retention, ad-revenue) into ML objectives, feature definitions, and success metrics, presenting findings to exec leadership every quarter.
- • Champion engineering excellence through code reviews, design docs, and brown-bag sessions; mentor junior engineers and set coding standards that balance velocity, readability, and operational resilience.
- • Contribute to open-source communities and internal tech talks, keeping the team at the forefront of advances in recommender systems, causal inference, and responsible AI.
🎯 Requirements
- • 4+ years production experience building large-scale ML systems in Python, Java, Golang, or Scala.
- • Expert-level proficiency with Spark, Kubernetes, TensorFlow/PyTorch, and distributed model serving.
- • Proven track record designing low-latency, high-throughput services with caching, load-balancing, and autoscaling.
- • Deep understanding of A/B testing, causal inference, and online experimentation best practices.
- • Hands-on experience with MLOps tools such as MLflow, Kubeflow, Airflow, feature stores, and drift monitoring.
- • Bachelor’s or Master’s degree in Computer Science, Engineering, or related quantitative field.
🏖️ Benefits
- • Competitive compensation plus annual performance bonus and ESOPs.
- • Comprehensive health insurance for employee and family, including mental-wellness programs.
- • Flexible PTO, quarterly recharge weeks, and hybrid work options to maintain work-life balance.
- • Annual learning stipend, global conference sponsorships, and internal hackathons to fuel continuous growth.
Skills & Technologies
Python
Java
Go
Scala
React
Senior
Onsite
Degree Required
About Jiostar Technologies Private Limited
Jiostar Technologies Private Limited is an Indian technology company delivering enterprise-grade digital transformation solutions. The firm specializes in cloud infrastructure, data analytics, cybersecurity, and AI-driven business applications for telecom, finance, and retail sectors. Its unified platform integrates IoT, edge computing, and API management to streamline operations and enhance customer engagement. Operating from Mumbai and Bengaluru, Jiostar serves large corporations seeking scalable, secure, and compliant technology stacks across India and emerging markets.
