Job Description- Architect the overall framework and infrastructure for GenAI products like search interfaces, bots, summarizers, etc. Develop and implement techniques to optimize model performance to meet specific product goals.
- Collaborate closely with product management and engineering leads to align on technical roadmap. Guide engineering teams to effectively leverage LLM capabilities in product implementations.
- Establish protocols and systems for building fair, accountable and transparent LLM based applications. Lead efforts to proactively assess and mitigate risks due to model biases or failures.
- Implement robust feedback pipelines, monitoring and corrections to ensure model safety.
- Design and oversee curation of high-quality datasets tailored for LLM training for each product. Build data science pipelines from feature generation, data visualization and models' evaluation. Design the solution, build initial code and provide documentation with ways of working to maximize time to value and re-usability.
- Communicate clearly and effectively to technical and non-technical audiences, verbally and visually, to create understanding, engagement, and buy-in. Contribute novel research and analyses to leading academic conferences and journals.
- Identify trends and opportunities to drive innovation, both in what we do and how we do it. Evaluate new data science, machine learning, and AI technologies and tools that can boost team performance, innovation, and business value. Proactively analyze latest developments in large language models to deeply understand model capabilities, limitations, and best practices. Develop techniques to continually improve language understanding and model training.
- Embody the values and passions that characterize Harvard Business School, with empathy to engage with colleagues from a wide range of backgrounds.
- Mentor and develop junior data scientists in state-of-the-art GenAI methods.
- Settle technical vision and lead initiatives to accelerate product impact through cutting edge LLM innovations.
- Complete other responsibilities as assigned.
Qualifications- Minimum of nine years' post-secondary education or relevant work experience.
- Advanced degree in mathematics, physics, computer science, engineering, statistics, or an equivalent technical discipline desired.
- Minimum of three years' experience in developing machine learning models with a track record of creating meaningful business impact and working with multiple stakeholders.
- Minimum of five years' experience with Python.
- Minimum of three years' experience building production NLP and deep learning models using PyTorch/Tensorflow, along with using large language model architectures(BERT, GPT-3 etc.)
- Experience building advanced workflows such as retrieval augmented generation, model chaining, dynamic prompting, PEFT/SFT, etc. using LangChain and similar tools.
- Experience in establishing model guardrails and developing bias detection and mitigation techniques for AI applications.
- Proficiency with various prompting techniques, with a clear understanding of trade-offs between prompting and finetuning.
- Experience with fine tuning embedding models and tuning vector databases to improve performance of semantic search and retrieval systems.
- Deep understanding of underlying fundamentals such as Transformers, Self-Attention mechanisms that form the theoretical foundation of LLMs.
- Experience with cloud computing platforms and tools (AWS, GCP, or other).
- Experience in operationalizing end-to-end machine learning applications.
How to ApplyBe part of a collaborative, fast-paced team at the forefront of innovation and technology advancements. Not only will you enjoy your work life at Relevance Lab, you'll also have the opportunity to grow your skills and career. If you are passionate about driving results, we'd love to talk with you.
