Machine Learning Engineer

Are you a builder who wants to solve hard problems in AI for Science? Atinary Technologies is revolutionizing R&D and Materials Discovery. Our platform, SDLabs, empowers researchers to accelerate scientific breakthroughs using AI-driven experimentation. We are looking for a Machine Learning Engineer to join the AI & Data Engineering (AIDEn) team and help us build the next generation of our ML product.

In this role, you will engineer the systems that translate Bayesian Optimization research and Generative AI concepts into scalable, production-grade software. You will serve as the bridge between our Research Scientists and our core platform, helping us scale from “algorithm” to “industry-changing product.”

Atinary Technologies is a fast-growing AI startup revolutionizing R&D through cutting-edge machine learning. We empower researchers and enterprises to accelerate discovery with our AI-driven platform. Our team values innovation, collaboration, and agility.

As a key member of the AIDEn team, you will take ownership of the engineering behind our ML capabilities. You will work on our model-serving infrastructure, our core Adaptive Experimentation library, and our evolving data processing architecture. This is a role for a versatile engineer with a systems mindset who is proficient in Python, cares about code quality, and is fascinated by the intersection of probabilistic machine learning and backend software engineering.

• Collaborate with the Research team to refactor, optimize, and product ionize experimental code (e.g., transitioning research prototypes into robust production features).
• Design and maintain the architecture for serving diverse ML models ensuring low latency and high availability.
• Build the backend systems required to ingest, validate, and store complex experimental data. You will help architect pipelines using tools like (MLFlow, AWS Sagemaker, AWS Batch, Ray) to support large-scale compute and inference workflows.
• Automate the lifecycle of our models. Ensure that when our researchers prototype a new optimization strategy, it can be deployed seamlessly using CI/CD, containerization, and orchestration.
• Contribute to the integration of Generative AI into our platform (using Lang Smith / Lang Chain), focusing on RAG, fine-tuning, and reliable structured outputs for scientific workflows.
• Work with Site Reliability Engineering (SRE) to ensure our ML services are cost-optimized, monitored, and resilient.

Must-have:

• 3+ years of experience in Machine Learning Engineering or Backend Engineering with a heavy data focus.
• Python Expertise:
  You write clean, type-hinted, testable code. You are familiar with modern standards.
• Data Proficiency & Tooling:
  Demonstrated proficiency in data handling, including experience with relevant tools such as Pandas, Num Py, scikit-learn, etc.
• ML Framework fluency:
  Comfortable with PyTorch, Tensorflow, Keras or equivalent.
• Infrastructure Mindset:
  Experience with containerization (Docker)

Nice-to-have:

• A genuine interest in how AI can accelerate scientific discovery.
• Bayesian Optimization:
  Experience with, or a strong desire to learn, BoTorch and GPyTorch.
• LLM Stack:
  Familiarity with the modern GenAI stack (Lang Chain, Lang Smith, Vector Databases).
• Product-Centric Mindset:
  You think beyond the model. You care about inference latency, API design, and how ML features integrate into the full-stack application to deliver a seamless user experience.
• Cloud Environment
  
  Experience:
  
  Experience with cloud services and environments (AWS preferred).
• Data Engineering
  
  Experience:
  
  Experience handling data pipelines and distributed compute frameworks (e.g., Ray, Dask, or Spark).

WHY YOU WOULD LOVE WORKING WITH US:

• Industry Impact:
  Your code will directly enable the discovery of new materials, drugs, and sustainable technologies.
• Independence & Versatility:
  We value engineers who can see the big picture. You won’t be siloed into a narrow task; instead, you will have the agency to design end-to-end solutions—whether that means optimizing a mathematical kernel or architecting a distributed computing system.
• Technical Challenge:
  Yo…