AI Engineering Intern - LLM Evaluations

Who We Are

Bynd is redefining financial intelligence through advanced AI, transforming how leading investment banks, private equity firms, and equity researchers globally analyze and act upon critical information. Our founding team includes a Partner from Apollo ($750B AUM) and AI engineers from UIUC, IIT, and other top-tier institutions. Operating as both a research lab and a product company, we build cutting-edge retrieval systems and AI-driven workflow automation for knowledge-intensive financial tasks.

Role Overview

As an AI Intern at Bynd, you’ll work at the intersection of cutting-edge GenAI systems and rigorous classical ML evaluation methodologies. Your primary responsibility will be to build and refine evaluation pipelines for our existing AI-driven financial intelligence systems. You’ll collaborate closely with the founding team and top financial domain experts to ensure our models are not only powerful—but measurable, explainable, and reliable.

If you’re excited by the idea of working hands-on with state-of-the-art LLMs, experimenting with RAG systems, and building frameworks that make AI outputs trustworthy and actionable, this role is made for you.

Responsibilities

• Design, implement, and iterate on evaluation pipelines for existing AI/ML systems, particularly GenAI-based and RAG-based architectures.

• Develop test sets, metrics, and validation frameworks aligned with financial use cases.

• Analyze model performance (both quantitative and qualitative) to uncover insights, gaps, and opportunities for improvement.

• Work alongside full-stack and ML engineers to integrate evaluation systems into CI/CD workflows.

• Assist in data collection, benchmark tasks, and A/B testing setups for LLM responses.

• Stay up-to-date with academic and industry advancements in evaluation frameworks, prompt testing, and trustworthy AI.

Preferred:

• Prior hands-on experience with GenAI systems (e.g., OpenAI, Claude, Mistral, etc.), including prompt design and retrieval-augmented generation (RAG).

• Solid understanding of classical ML concepts like training-validation splits, overfitting, data leakage, and cross-validation.

• Familiarity with tools such as Weights & Biases, LangSmith, or custom logging/benchmarking suites.

• Comfort with Python, evaluation libraries (e.g., sklearn, evaluate, bert-score, BLEU/ROUGE, etc.), and backend integration.

• Experience working with unstructured financial data (PDFs, tables, earnings reports, etc.) is a massive plus.

What We’re Looking For

We’re looking for a fast learner with deep intellectual curiosity and strong fundamentals. You should be comfortable reasoning through ambiguity, rapidly testing hypotheses, and communicating technical decisions with clarity. You’re someone who thinks not just about building intelligent systems—but about how we measure intelligence meaningfully.

This is an opportunity to work closely with a high-caliber founding team and ship impactful systems used by decision-makers at global financial institutions. If you’re passionate about building AI that works and works reliably, come build with us.