Member of Technical Staff, Training Infra at Inception (San Francisco Bay Area, CA)

Inception· San Francisco Bay Area, CA· $200k–$350k·1d ago

Role details

Salary

$200k–$350k

Work type

Onsite

Employment

Full-Time

Equity

Yes

Job description

The Role

We're looking for engineers and scientists to design, optimize, and maintain the core systems that enable scalable, efficient training of LLM. Your goal is to make experimentation and training at Inception fast and reliable so our team can focus on science, not system bottlenecks.

Key Responsibilities

Design, implement, and optimize distributed training systems that scale across thousands of GPUs and nodes.
Develop high-performance optimizations to maximize throughput and efficiency.
Develop reusable frameworks and libraries to improve training reproducibility, reliability, and scalability for new model architectures.

Qualifications

BS/MS/PhD in Computer Science, Engineering, or a related field (or equivalent experience).
Understanding of ML frameworks (PyTorch, TensorFlow) from a systems perspective.
Strong engineering skills — ability to contribute performant, maintainable code and debug in complex codebases.
Proficiency in Python and at least one systems programming language (C++/Rust/Go).
Experience with containerization (Docker), orchestration (Kubernetes), and CI/CD pipelines.

Preferred Skills

Experience building and maintaining large-scale language models with tens of billions of parameters or more.
Experience with ML workflow orchestration tools (Kubeflow, Airflow).
Background in performance optimization and profiling of ML systems (Prometheus, Grafana, OpenTelemetry).
Familiarity with distributed frameworks such as PyTorch/XLA, DeepSpeed, Megatron-LM.

Why work at Inception

Culture: A tight‑knit team of scientists, engineers, and builders focused on shipping frontier AI research. Emphasis on innovation, speed, and real‑world impact.
Work policy: In‑office for most roles (Palo Alto HQ); a Marketing Intern position is listed as remote. Typical time on‑site is full‑time in the office.
Notable perks / engineering culture: Opportunity to work on cutting‑edge diffusion LLMs from scratch, contribute to foundational research (papers published), and collaborate with alumni from top AI labs. Roles span from kernel engineering to RL infrastructure and product management.
Growth: Rapidly scaling headcount (+187% YoY) with open positions across multiple disciplines, offering strong career progression in a high‑visibility startup.

About Inception

Company Overview

One-liner: Inception is an AI research and product company building diffusion-based language models (dLLMs) for production applications, offering 5x faster inference than traditional autoregressive LLMs.
Entity Type: Private (Seed-stage; $57M total funding)
Headquarters: Palo Alto, California, United States
Founded: Not publicly available (earliest funding round dated March 2025)
Founders: Stefano Ermon (CEO), Aditya Grover (Co-Founder & CTO), Volodymyr Kuleshov (Co-Founder)

Core Business

Primary industry: Artificial intelligence research and infrastructure; large language models (LLMs)
Target customers: B2B developers, enterprises, and AI application builders requiring low-latency, high-volume LLM inference for multi-step agents and real-time products.
Mission or purpose: “We’re building the next generation of LLMs” – enabling intelligent, responsive, and scalable AI applications through diffusion-based architectures.

Products & Services

Mercury (dLLM family): The world’s first commercially available family of diffusion large language models. Generates output via a coarse-to-fine iterative refinement process over a small number of steps, achieving 5x greater speed than autoregressive models while maintaining best-in-class quality. Seamlessly integrates into existing LLM workflows.
Research technologies: d1 Reasoning, Discrete Diffusion Guidance, Direct Preference Optimization, Flash Attention, Decision Transformers, and Diffusion Models – foundational contributions that underpin the company’s product.

Market Standing

Valuation / Market Cap: Not disclosed
Key Metric: Total funding of $57M across two seed rounds (2025)
Notable Investors / Partners: Menlo Ventures (led the $50M seed round), Amazon Web Services (non‑equity assistance), Eric Schmidt (individual investor)
Growth Signals: Headcount grew 187.5% year‑over‑year to 39 employees; LinkedIn followers increased 240% yearly to ~11,800; active job postings across AI systems, engineering, research, product, and marketing; talent sourced from Google DeepMind, Stanford SAIL, Meta, Amazon, and Pixxel.

Competitive Advantages

Diffusion paradigm: Non‑autoregressive generation that allows parallel refinement and revision during inference, breaking the sequential token‑by‑token bottleneck of conventional LLMs.
Speed‑cost curve: Delivers 5x faster output with a fundamentally different cost structure, critical for real‑time, multi‑step agent workflows.
Founding team pedigree: Co‑founders have pioneered breakthroughs in model architectures (Flash Attention, Direct Preference Optimization) and have deep experience turning research into production systems at scale.

Strategic Focus

Scaling Mercury adoption by integrating dLLMs into existing LLM pipelines and targeting high‑volume, latency‑sensitive applications.
Advancing research in diffusion‑based reasoning, reinforcement learning for language models, and training/serving infrastructure.
Expanding the team across AI systems (kernels, inference, training infra), product engineering, and research to accelerate product‑market fit.

Sources