Python Developer

Moleculyst Ventures Private Limited is an AI startup building software for detecting and analyzing hallucinations in large language models (LLMs).

Our work focuses on understanding when and why language models generate unreliable outputs and developing systems that can identify these failures during model inference. We work with open-weight language models and analyze model outputs, internal computations, and inference-time signals to build reliable hallucination detection and monitoring systems.

The broader goal is to build infrastructure that makes AI systems more reliable, observable, and trustworthy for real-world deployment.

We are looking for a Python Developer to help build the software infrastructure behind our hallucination detection platform.

You will work on systems for running language models, capturing and processing inference-time data, building data pipelines, implementing backend services, and storing and analyzing large volumes of model-generated information.

You will collaborate closely with AI/ML researchers to turn experimental methods into reliable and scalable software.

• Design, build, and maintain clean, modular, and testable Python systems.
• Build infrastructure for running and evaluating open-weight large language models.
• Develop pipelines for capturing, processing, filtering, and storing inference-time data.
• Build APIs and backend services for model evaluation and hallucination detection.
• Work with large numerical datasets, model activations, tensors, metadata, prompts, and generated outputs.
• Integrate databases and storage systems for efficient retrieval and analysis.
• Write unit and integration tests to ensure correctness and reproducibility.
• Profile and optimize performance-critical Python code.
• Collaborate with AI/ML researchers to convert experimental research code into robust software.
• Document system architecture, APIs, data formats, and engineering decisions.

• Strong Python programming skills.
• Strong understanding of object-oriented programming, data structures, and software design.
• Ability to write clean, modular, maintainable, and well-tested code.
• Experience with NumPy and PyTorch.
• Familiarity with backend development and building APIs.
• Experience with SQL and/or NoSQL databases.
• Understanding of Git, testing, debugging, and standard software-development practices.
• Experience processing large or complex datasets.
• Comfortable working with Linux and command-line development environments.
• Ability to independently investigate technical problems and implement solutions.
• Strong communication skills and ability to work effectively in a remote team.

• Experience with PyTorch internals or transformer-based language models.
• Familiarity with Hugging Face Transformers or other open-weight LLM tooling.
• Experience running or serving LLMs.
• Understanding of GPU computing, CUDA, or performance optimization.
• Experience with Docker and cloud infrastructure.
• Familiarity with model inference, forward hooks, activations, attention, KV caches, or other transformer internals.
• Experience building ML experimentation or evaluation infrastructure.

A typical system you may help build looks like:

Prompt → LLM Inference → Internal/Output Data Capture → Signal Processing → Hallucination Detection → Storage/API

The role sits at the intersection of Python engineering, ML infrastructure, and LLM interpretability/reliability.

• This is a full-time remote role requiring ownership, independent problem-solving, and close collaboration with the research team.