X Logo
X· HQ in United States

Sr. Software Engineer, Training/Inference (C++)

Location: Palo Alto, CA · Category: Engineering & Development · Seniority: Senior · Posted: November 6, 2026 · Verified: Nov 21, 2026

Role Overview & Description

This role designs and optimizes large-scale model serving systems end-to-end, owning distributed infrastructure including global KV cache, continuous batching, load balancing, and auto-scaling. The engineer performs deep low-level optimizations such as GPU kernels, quantization, speculative decoding, and tail latency reduction, requiring 5+ years of low-level system programming in C/C++ or Rust and experience with GPU inference engines.
Hiring Telemetry · Expansion Signal

Track team expansion and market signals for X

Polaris7 tracks open roles, strategic market movements, and M&A developments from verified primary sources. Receive structured Watchlist briefings on X and benchmark against direct competitors.

Source & Transparency: This position was structured directly from the publicly accessible career portal of X. Details on the employer’s original job posting are authoritative. Polaris7 charges no placement fees and conducts no candidate screening.
Company IntelligenceVerified Profile

Operator of X, a social platform monetised through ads, subscriptions and API access.

Headquarters: United States

Similar Positions in DACH

Related roles across advertising, media, and data

Verified 2w agoPalo Alto, CAEngineering & DevelopmentSenior

This role architects and implements scalable distributed infrastructure for model serving, including global KV cache, continuous batching, load balancing, and auto-scaling. The engineer will optimize latency and throughput of model inference, build high-concurrency serving systems, and accelerate research on scaling test-time compute. Requires 5+ years of low-level system programming in C/C++ or Rust and experience with GPU inference engines like vLLM or TensorRT-LLM.

Re-check scheduledSan FranciscoEngineering & DevelopmentSenior

This Kernel Engineer role at an AI company focuses on making machine learning training and inference faster, more efficient, and more reliable at the hardware-software boundary. The engineer will specialize in GPU/accelerator kernels, ML compilers, or framework/runtime and inference performance, optimizing compute primitives, memory layouts, and code generation. The role requires deep implementation-level expertise in stacks like CUDA, Triton, XLA, or PyTorch/JAX backends.

Re-check scheduledSan FranciscoEngineering & DevelopmentSenior

Optimizes inference and training performance for large generative world models, profiling and eliminating bottlenecks in kernels, serving paths, and GPU utilization. Writes and tunes GPU kernels (CUDA/Triton), implements low-precision execution (FP8/INT8), and builds performance models and observability frameworks. Requires deep GPU programming experience, strong performance engineering foundations, and proficiency in Python with C++/CUDA capabilities.

Re-check scheduledSan FranciscoEngineering & DevelopmentSenior

This role focuses on optimizing inference performance for large-scale AI models, including model serving, latency reduction, and throughput maximization. The engineer will work on inference engines, batching strategies, and hardware acceleration to support the company's open superintelligence platform. Key challenges include building efficient inference pipelines and integrating them with distributed training and deployment systems.

More Positions at X

3 more open roles

Verified todaySouthaven, MS; Memphis, TNEngineering & DevelopmentLead

This role leads operations engineering for construction and site efficiency at SpaceXSI's data centers, focusing on process improvement, metrics dashboards, and automation. The engineer will use SQL and Python for data analysis, implement tools like Power BI and JIRA, and drive continuous improvement across maintenance and construction teams. Requires 7+ years of operations experience and 2+ years leading teams.

Verified todayRemote, InternationalRemoteData, AI & AnalyticsMid-LevelUSD45 – USD75

This role involves creating high-quality data annotations and inputs for biological contexts to train the Grok AI model. The AI Tutor will evaluate model outputs, create training data, and help Grok reason correctly from experimental evidence in areas like molecular biology, genomics, and protein science. The position requires a Master's or PhD in a biology field and expertise in specialized subdomains.

Verified todayRemote, InternationalRemoteData, AI & AnalyticsMid-LevelUSD45 – USD100

This role involves creating and annotating high-quality medical data to train the Grok AI model, spanning text, voice, and video. The AI Tutor will evaluate model outputs, identify errors in clinical reasoning, and capture how clinicians think and communicate with patients. The position requires a medical degree (MD, DO, MBBS, etc.) or a Master's/PhD in a health-related field with 2+ years of clinical or research experience.

Start with a company or market. Track what changes.

Create your Polaris7 workspace to explore extended company profiles, organize relevant players into Watchlists, and receive focused briefings when material changes occur.