Research Scientist Intern (Python)

Collinear; Bengaluru; On-site; INR 900,000 - 1,500,000; Full time

Job summary

Collinear is seeking a research-oriented engineer to explore foundation models, designing reproducible experiments and scalable infrastructure. You will work in a Python-centric, CLI-driven environment to build end-to-end pipelines that translate research insights into production-ready systems.

The role emphasizes rigorous evaluation, data-supported findings, and a strong ownership mindset to drive measurable outcomes in a challenging technical domain.

Qualifications

• Degree in Computer Science, Mathematics, Physics, or related technical discipline; equivalent open-source work or experience accepted.
• Strong software engineering skills to design and implement dependable systems.
• Proficiency in Python and comfort with CLI-based environments.
• Deep understanding of foundation models, including training, evaluation, and tuning approaches.
• Evidence of reproducible, data-supported research outcomes.
• Ownership mindset with focus on measurable outcomes.

Responsibilities

• Conduct research on foundation models, exploring construction, assessment, and enhancement methods.
• Design and execute technical experiments with reproducible methodologies and quantitative validation.
• Build and maintain infrastructure that is robust, scalable, and production-ready.
• Work within a Python-centric, command-line-driven workflow.
• Balance exploratory research with delivering working systems.

Skills

Python

CLI proficiency

Software architecture

Foundation models knowledge

Reproducible research

Education

CS/Math/Physics degree

Tools

Python tooling

Job description

Batch: 2026/2027/2028.

About the role:

• Conduct research on foundation models, exploring their construction, assessment, and enhancement methods
• Design and execute technical experiments with emphasis on reproducible methodologies and quantitative validation
• Build and maintain infrastructure that is robust, scalable, and production-ready
• Work within a Python-centric, command-line-driven workflow
• Balance exploratory research initiatives with the practical requirements of delivering working systems

What we're looking for:

• Degree in Computer Science, Mathematics, Physics, or a related technical discipline; alternatively, demonstrated equivalent expertise through significant open-source work or professional background
• Solid grounding in software engineering with capability to architect and implement dependable systems
• Proficiency in Python and comfort operating in CLI-based environments
• Deep conceptual grasp of how foundation models function, including training, evaluation, and tuning approaches
• Track record of running research or technical investigations that yield reproducible, data-supported findings
• Genuine drive to identify and solve critical problems within intricate technical domains
• Ownership mindset with a strong focus on measurable outcomes

Desirable experience:

• Work with Reinforcement Learning, RLHF/RLAIF, simulation frameworks, or long-horizon agentic systems
• Published research at top-tier venues such as NeurIPS, ICLR, or ICML
• Contributions to widely-adopted open-source initiatives
• Fine-tuning or assessment of large-scale models against competitive benchmarks on publicly available evaluation sets