Senior Recommendation System Engineer

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI

Last checked: 3 hours agoStatus: Live

Job Summary

Design and refactor high-concurrency, low-latency recommendation systems that power multi-stage recall, ranking, and re-ranking. Build real-time feature pipelines and govern online/offline feature consistency using stream-batch convergence. Optimize large-scale vector retrieval with Faiss/Milvus/NSW, and deploy deep ranking models for low-latency inference. Ensure global performance with SLOs, distributed tracing (Jaeger/Prometheus), and experimentation infrastructure for faster iteration.

Location: Kuala Lumpur

Workplace: Onsite

Employment Type: Full time

Job Function: Solutions Engineering & Sales Engineering

Seniority: Mid level

Key Responsibilities

•Own development and refactoring of high-concurrency, low-latency recommendation serving engines across multi-channel recall (two-tower/collaborative filtering/ANN vector retrieval), coarse ranking, fine ranking, and re-ranking.
•Implement compute tiering and strategy mechanisms, including dynamic compute trimming and degradation for UI personalization and global strategy dispatch under extreme traffic spikes.
•Build high-throughput, low-latency real-time feature pipelines on Kafka/Flink and contribute to unified online/offline feature storage to address inconsistency and time-travel leakage.
•Construct and optimize large-scale vector retrieval systems (Faiss/Milvus/NSW), including index tuning to achieve low P99 retrieval latency and build unified embedding/inverted-index services across heterogeneous sources.
•Ensure high availability and SLO targets, build global distributed tracing and monitoring (Jaeger/Prometheus), and contribute to A/B experimentation platform upgrades with CUPED and sequential testing.

Pay and Benefits

Perks:Learning Budget

Key Requirements

•5+ years of recommendation system engineering at consumer-scale internet companies, including deep participation in or leadership of architecture refactoring or launches of real-time recommendation systems serving tens of millions of users.
•Strong engineering fundamentals; proficient in at least one of Go/Java/C++ (Go preferred) and familiar with online inference and deployment optimization using PyTorch/TensorFlow.
•Hands-on experience with the Spark/Flink/Kafka big-data stack, solving stream-computing latency/backlog issues, and proficient in Milvus/Faiss cluster deployment and tuning.
•Deep understanding of computational complexity and online bottlenecks for core recommendation algorithms (collaborative filtering, two-tower recall, MMoE/PLE multi-objective optimization) and ability to work with algorithm teams.
•Proven experience designing and developing recommendation platforms, feature platforms, experimentation platforms, or high-performance RPC frameworks with strong system design capability.

Experience:5+ yearsConsumer internetRecommendation systemsBig dataReal-time systems

Skills:Problem-solvingSystem designEngineering judgment

Languages:English

Tech Stack:GoJavaC++PyTorchTensorFlowSparkFlinkKafkaMilvusFaissNSWKafka/FlinkTwo-towerCollaborative filteringANNDINSIMMMoEPLEJaeger

Company Brief

Bybit

Website

Loading job details...

Preparing the role view and application actions.

Bybit

3 weeks ago

Senior Recommendation System Engineer

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI

Last checked: 3 hours agoStatus: Live

Job Summary

Location: Kuala Lumpur

Workplace: Onsite

Employment Type: Full time

Job Function: Solutions Engineering & Sales Engineering

Seniority: Mid level

Key Responsibilities

•Own development and refactoring of high-concurrency, low-latency recommendation serving engines across multi-channel recall (two-tower/collaborative filtering/ANN vector retrieval), coarse ranking, fine ranking, and re-ranking.
•Implement compute tiering and strategy mechanisms, including dynamic compute trimming and degradation for UI personalization and global strategy dispatch under extreme traffic spikes.
•Build high-throughput, low-latency real-time feature pipelines on Kafka/Flink and contribute to unified online/offline feature storage to address inconsistency and time-travel leakage.
•Construct and optimize large-scale vector retrieval systems (Faiss/Milvus/NSW), including index tuning to achieve low P99 retrieval latency and build unified embedding/inverted-index services across heterogeneous sources.
•Ensure high availability and SLO targets, build global distributed tracing and monitoring (Jaeger/Prometheus), and contribute to A/B experimentation platform upgrades with CUPED and sequential testing.

Pay and Benefits

Perks:Learning Budget

Key Requirements

•5+ years of recommendation system engineering at consumer-scale internet companies, including deep participation in or leadership of architecture refactoring or launches of real-time recommendation systems serving tens of millions of users.
•Strong engineering fundamentals; proficient in at least one of Go/Java/C++ (Go preferred) and familiar with online inference and deployment optimization using PyTorch/TensorFlow.
•Hands-on experience with the Spark/Flink/Kafka big-data stack, solving stream-computing latency/backlog issues, and proficient in Milvus/Faiss cluster deployment and tuning.
•Deep understanding of computational complexity and online bottlenecks for core recommendation algorithms (collaborative filtering, two-tower recall, MMoE/PLE multi-objective optimization) and ability to work with algorithm teams.
•Proven experience designing and developing recommendation platforms, feature platforms, experimentation platforms, or high-performance RPC frameworks with strong system design capability.

Experience:5+ yearsConsumer internetRecommendation systemsBig dataReal-time systems

Skills:Problem-solvingSystem designEngineering judgment

Languages:English

Tech Stack:GoJavaC++PyTorchTensorFlowSparkFlinkKafkaMilvusFaissNSWKafka/FlinkTwo-towerCollaborative filteringANNDINSIMMMoEPLEJaeger

Company Brief

Bybit

Website