Lead Data Science Researcher

So, what’s the role all about?

How will you make an impact?

• Lead end‑to‑end research initiatives across NLP, Vision, and multimodal modeling, with a strong focus on LLM‑based and agentic‑AI systems.

• Architect, prototype, and evaluate singleagent and multiagent systems, including planning, tool use, memory, and orchestration.

• Establish and own best practices for safe, controllable, and scalable AI agents, including evaluation frameworks, guardrails, fallback strategies, and observability.

• Act as a technical authority on LLM and agentic systems, guiding architectural decisions, evaluation strategies, and engineering tradeoffs.

• Define rigorous offline and online evaluation strategies (KPIs, A/B testing, cost/performance tradeoffs) grounded in realworld constraints.

• Deliver select research components at production quality and partner closely with productization teams to harden and deploy endtoend solutions.

• Mentor researchers and data scientists, raising the bar for technical rigor, engineering quality, and applied research impact.

• Communicate complex findings and risks clearly to crossfunctional stakeholders and leadership.

• Stay at the forefront of AI research, contributing to the team’s agentic‑AI roadmap and long‑term research vision and help set teamwide standards and guidelines.

Have you got what it takes?

• Skills-first profile with proven, hands‑on impact in applied AI. (Formal degrees welcome but not required.)

• Strong Demonstrated experience building production‑grade AI agents that perform multi‑step reasoning and tool‑based actions (e.g., tool invocation, planning, memory).

• Mandatory: Experience with agent frameworks/orchestration layers or custom agent runtimes (e.g., LangGraph/LangChain, semantic routers, workflow engines, or in‑house frameworks).

• Strong practical expertise with LLMs (AWS Bedrock or similar platforms), including evaluation, prompt/program design, and safety patterns.

• Strategic problem-solving leadership: You proactively shape ambiguous business questions into well-defined analytical goals, challenge underlying assumptions, and ensure the work is focused on the problems with the highest impact.

• Bar‑setting analytical rigor—applied pragmatically: You anticipate bias, confounders, and risks of misinterpretation early, apply the right level of methodological rigor for the decision at hand, and help others distinguish between “theoretically perfect” and “fit for purpose.”

• Efficient, scalable thinking: You balance depth with speed, favor simple and robust solutions over unnecessary complexity, turn one‑off analyses into reusable insights, and help the organization avoid reinventing or over‑engineering solutions.

• Track record of translating research into reliable systems in partnership with engineering/product teams.

• Proficiency in Python and modern ML/DL libraries; strong AI engineering skills (clean architecture, testing, CI/CD, observability), as well as familiarity with HuggingFace or similar model ecosystems and open‑source tooling.

• Experience applying GenAI to DS workflows (LLM‑as‑a‑judge, synthetic data generation, weak labeling, automated eval).

• Excellent communication and mentoring skills; fluency in English.

You will have an advantage if you also have:

• Skills & Tags

  pythongoawsaidataproductdesign