Post-Training and Reinforcement Learning Engineer

About the company

SustainabilityEconomics.ai is a Singapore- and US-based company with a large development center in Bengaluru. The company is building AI agents that are paid on the results they deliver. The first is Clem.ai, an Energy-Mix Agent that plans, contracts and operates the power supply mix for large electricity consumers such as data centers. Its objective is power that is lower in cost than grid supply, at least as reliable, and carbon-free around the clock.

The company does not pre-train models. It takes open models and post-trains them on narrow energy tasks where a program can verify the result. Solvers, calculators and simulators perform the arithmetic, and the models are trained to use them correctly.

The position

We are looking for a Post-Training and Reinforcement Learning Engineer to join the platform team in Bengaluru. The engineer will help train, evaluate and improve the open models behind Clem.ai.

This is an early-career position reporting to the Post-Training Lead. The engineer will work alongside senior post-training engineers, applied AI engineers and the company’s energy experts. An independent assurance team approves every model release.

Scope of work

• Prepare, clean and verify training data for tasks such as bill and invoice extraction, contract clause tagging and optimization template completion.
• Run supervised fine-tuning of open models using low-rank adapters (LoRA) on company GPU infrastructure.
• Develop reinforcement learning environments and reward functions in which a validator, solver, calculator or simulator scores the model output.
• Evaluate every model before and after training on the company’s energy-domain benchmark, and report the results accurately.
• Convert operator corrections and overrides from live operation into new training examples and evaluation cases.
• Generate and validate synthetic data where real examples are limited.
• Support the distillation of larger models into smaller, lower-cost models.
• Maintain reproducible experiments, including configurations, data versions, run logs and written summaries.
• Use coding agents for first drafts of code, tests and data preparation, and review all output before use.
• Comply with data governance requirements, including keeping customer data used for per-site tuning within that customer’s sandbox.

Required qualifications

• Bachelor’s or master’s degree in computer science, electrical engineering, mathematics or a related discipline.
• Up to two years of relevant experience, including internships and research projects.
• Strong proficiency in Python and working knowledge of PyTorch.
• Demonstrated experience fine-tuning at least one open language model, for example with Hugging Face Transformers and LoRA.
• Sound understanding of machine learning fundamentals, including data splits, overfitting and evaluation metrics.
• Working knowledge of reinforcement learning fundamentals, including policies, rewards and environments.
• Proficiency with Git, Linux and GPU-based training jobs.
• Rigor in data handling and evaluation, and the ability to report results objectively.
• Clear written communication in English.

Prior experience in the energy sector is not required.

Preferred background

• Practical experience with preference-based or reinforcement learning methods for language models, such as DPO or GRPO-style training.
• Familiarity with reinforcement learning algorithms such as PPO or SAC, and tools such as Gymnasium.
• Experience writing reward functions or automated graders.
• Experience with distillation or synthetic data generation.
• Experience with structured outputs and tool calling in language models.
• Exposure to optimization solvers or time-series forecasting.
• Interest in power markets, renewable energy or data center operations.
• Open-source contributions, competition results or published research.

Personal attributes

• Leadership skills, including the ability to take ownership of work and guide it to completion.
• Strong self-motivation and the initiative to work with limited supervision.
• Eagerness to learn new methods and a new domain quickly.
• Effective teamwork with engineers, energy experts and the assurance team.
• Accountability for results and integrity in reporting them.
• Ability to manage time and deliver to agreed deadlines.

Location and application

The position is based in Bengaluru.

Interested candidates should apply directly through LinkedIn, including their CV and a link to one fine-tuning or reinforcement learning project.