| Job Position | Company | Posted | Location | Salary | Tags |
|---|---|---|---|---|---|
Nexus | San Francisco, CA, United States | $126k - $127k | |||
Nexus | San Francisco, CA, United States | $98k - $120k | |||
Nexus | San Francisco, CA, United States | $98k - $106k | |||
Raiku | London, United Kingdom | $90k - $112k | |||
| Learn job-ready web3 skills on your schedule with 1-on-1 support & get a job, or your money back. | | by Metana Bootcamp Info | |||
Polygon Labs | United States | $72k - $100k | |||
Nexus | San Francisco, CA, United States | $105k - $180k | |||
Aztec | Remote | $72k - $77k | |||
Aztec | Remote | $81k - $95k | |||
Logos | New York, NY, United States | $63k - $68k | |||
Vac | New York, NY, United States | $63k - $68k | |||
Starkware | Remote | $175k - $240k | |||
Sphere | United States | $103k - $195k | |||
Sphere | United States | $90k - $115k | |||
Polar Bear Systems. | United States | $175k - $240k | |||
Nexus | San Francisco, CA, United States | $81k - $150k |
About Nexus
Nexus is building a world supercomputer by leveraging the latest advancements in AI, cryptography, engineering, and science. Our team of world-leading experts is developing and deploying the Nexus Layer 1 blockchain and Nexus zkVM (zero-knowledge virtual machine) in support of our mission to enable the Verifiable Internet.
Nexus raised $25M in Series A funding from Lightspeed, Pantera, Dragonfly, SV Angel, and more.
We are headquartered in San Francisco, and this role will be in-person with the rest of the Nexus team.
AI Software Engineer, Inference
We’re looking for an AI Software Engineer focused on Inference to help us bring powerful AI models to life — fast, efficient, and at scale. This role is all about building the systems that deliver real-time predictions, keeping latency low and performance high. If you love optimizing ML workloads and making complex systems run smoothly in production, this one's for you.
At our startup, speed matters — not just in how our models perform, but in how quickly we learn, ship, and grow. You’ll be a core part of the engineering team, collaborating closely with researchers and product engineers to build scalable inference systems that support everything we do.
Responsibilities
Design and optimize lightning-fast inference pipelines for both real-time and batch predictions
Deploy and scale machine learning models in production across cloud and containerized environments
Leverage frameworks like TensorFlow Serving, TorchServe, or Triton to serve models at scale
Monitor performance in the wild — build tools to track model behavior, latency, and reliability
Work with researchers to productionize models, implement model compression, and make inference as efficient as possible
Solve problems fast — whether it’s a scaling bottleneck, a failed deployment, or a rogue latency spike
Build internal tools that streamline how we deploy and monitor inference workloads
RequirementsÂ
3+ years of experience in software engineering, preferably with exposure to ML systems in production
Strong skills in Python, Go, or Java, and a solid understanding of system performance fundamentals
Experience with containerization (Docker, Kubernetes) and deploying services in the cloud (AWS, GCP, or Azure)
Solid understanding of model serving architectures and techniques for optimizing latency and throughput
Comfort with performance tuning and profiling of ML model execution
A practical mindset and eagerness to own production systems from build to run
Embrace AI as a core part of how you work, think, and build.
Bonus Points
Experience with hardware acceleration for inference (GPUs, TPUs, etc.)
Familiarity with real-time data processing and streaming tools
Hands-on with edge deployment (mobile, embedded, etc.)
Contributions to open-source projects in model serving or ML infrastructure
Benefits
Competitive salary and generous equity compensation
Health insurance for employees and their dependents
Daily lunch and dinner provided at SF headquarters
Company-paid travel to events and conferences
Nexus is committed to diversity in our workforce and is proud to be an Equal Opportunity Employer (EEO).
What is Zero-knowledge?
Zero-knowledge is a concept in cryptography that allows two parties to exchange information without revealing any additional information beyond what is necessary to prove a particular fact
In other words, zero-knowledge is a way of proving something without actually revealing any details about the proof
Here are some examples of zero-knowledge:
- Password authentication: When you enter your password to log into an online account, the server doesn't actually know your password. Instead, it checks to see if the hash of your password matches the stored hash in its database. This is a form of zero-knowledge because the server doesn't know your actual password, just the hash that proves you know the correct password.
- Sudoku puzzles: Suppose you want to prove to someone that you've solved a particularly difficult Sudoku puzzle. You could do this by providing them with the completed puzzle, but that would reveal how you solved it. Instead, you could use a zero-knowledge proof where you demonstrate that you know the solution without actually revealing the solution itself.
- Bitcoin transactions: In a Bitcoin transaction, you prove that you have ownership of a certain amount of Bitcoin without revealing your private key. This is done using a zero-knowledge proof called a Schnorr signature, which allows you to prove ownership of a specific transaction output without revealing the private key associated with that output.
- Secure messaging: In a secure messaging app, you can prove to your contacts that you have access to a shared secret without revealing the secret itself. This is done using a zero-knowledge proof, which allows you to prove that you have access to the secret without actually revealing what the secret is.