Job Description

Your Future Evolves Here

Evolent Health has a bold mission to change the health of the nation by changing the way health care is delivered. Our pursuit of this mission is the driving force that brings us to work each day. We believe in embracing new ideas, challenging ourselves and failing forward. We respect and celebrate individual talents and team wins. We have fun while working hard and Evolenteers often make a difference working in everything from scrubs to jeans.

Are we growing? Absolutely and Globally. In 2021 we grew our teams by almost 50% and continue to grow even more in 2022. Are we recognized as a company you are supported by for your career and growth, and a great place to work? Definitely. Evolent Health International (Pune, India) has been certified as \xe2\x80\x9cGreat Places to Work\xe2\x80\x9d in 2021. In 2020 and 2021 Evolent in the U.S. was both named Best Company for Women to Advance list by Parity.org and earned a perfect score on the Human Rights Campaign (HRC) Foundation\xe2\x80\x99s Corporate Equality Index (CEI). This index is the nation\'s foremost benchmarking survey and report measuring corporate policies and practices related to LGBTQ+ workplace equality.

We recognize employees that live our values, give back to our communities each year, and are champions for bringing our whole selves to work each day. If you\xe2\x80\x99re looking for a place where your work can be personally and professionally rewarding, don\xe2\x80\x99t just join a company with a mission. Join a mission with a company behind it.

What You\xe2\x80\x99ll Be Doing:

Position summary
As a Generative AI Engineer, you will be responsible for implementing and fine tuning cutting-edge generative AI models to solve complex problems, drive innovation and increase productivity across our organization. You will work closely with data scientists, software engineers, and product managers to design, build, and deploy GenAI-powered solutions that enhance our products and services.

These applications include but are not limited to advanced OCR (optical character recognition) using transformers/new features in GPT models, understanding, Summarization, Classification and QnA the contents of the medical documents (clinical notes, health policies) with help of large language models, and integrating the flow into existing product and services offerings.

We ingest claims, medical charts, etc. from providers containing unstructured data which will be used enhance productivity and intelligence across several specialty care division in evolent.

The outcome will be used dually for real-time operational processes with both automated and augmented human-based decision making (i.e., copilot tools) to reduce healthcare administrative costs. We work with all major cloud and big data vendors offerings including but not limited to (Azure, AWS, Google, IBM, etc.) to achieve AI goals in healthcare and support Evolent business.

Essential functions

Applying foundation models mostly Large Language Models (OpenAI GPT family of models) and best practices around Generative AI to solve some of complex problems in healthcare domain through our world class infrastructure on cloud. Also use other foundation models like computer vision models (ResNet, EfficientNet, YOLO) and Generative Models (DALL-e, GANs, VAEs) in augment the intelligence around administrative process and people working on these pdf/images/handwritten docs/forms

Optimize chunking strategies, embedding models (sentences transformers/text models), size of the LLMs and apply appropriate Model compression techniques and Attention Models.

Implement novel technique(s) to fine tune LLMs like instruction fine tuning, parameter-efficient fine-tuning methods, LoRA/QLoRA, Quantization algorithm, Reinforcement Learning with Human Feedback, and Retrieval-Augmented Generation (RAG) to improve prediction quality

Building LLMs pipeline using framework like LangChain, LlamaIndex, Semantic Kernel, OpenAI functions and apply state of art prompt engineering principle like persona, chain of thoughts etc to reduce LLMs hallucinations, understanding of zero-shot/few shot learning, and different LLM parameters like temperature, top-k, repeat penalty

Operationalizing and understanding best practices around different vector search technologies other than azure cognitive search like neo4j, elastic search, predibase, weaviate, cohere etc.

Experiment with different LLM models from LLMs providers (Google PaLM, Bedrock, AI21 labs, HuggingFace, Meta, Anthropic, NVIDIA), explore Multimodality of LLMs/ different features of LLMs to reduce manual efforts at different health plan operations and for clinical innovation. Track LLM leaderboard like HuggingFace Leaderboard for open LLM to understand latest models.

Build monitoring and performance tracking framework around LLMs and use LLMOps framework like mlflow, LangSmith and any latest technique for the same.

Rapid prototyping around latest trend like AI Agents, open source LLMs, cost effective LLMs i.e. HuggingGPT, Mistral to continuous find ways to improve performance

Deploying and creating LLMs plugin, copilots, SQL copilots to enhance the productivity of individuals across several vertical in the company and awareness of pair programming with AI.

Evaluate NLP outcomes with ROGUE/BLEU scores and evaluate LLMs based on various metrics like faithfulness, relevancy, harmfulness, perplexity, fluency, coherence, diversity and use framework like RAGAS. Understand and implement industry standard benchmark for LLM evaluation like \xe2\x80\x93 Abstraction and Reasoning Corpus (ARC) , HellaSWAG (commonsense), MMLU (Massive Multitask Language Understanding), TruthfulnessQA etc.

Collaborating with cross-functional teams to define AI project requirements and objectives, ensuring alignment with overall business goals.

Conducting research to stay up-to-date with the latest advancements in generative AI, machine learning, and deep learning techniques and identify opportunities to integrate them into our products and services.

Developing and maintaining AI pipelines, including data preprocessing, feature extraction, model training, and evaluation.

Developing clear and concise documentation, including technical specifications, user guides, and presentations, to communicate complex AI concepts to both technical and non-technical stakeholders.

Contributing to the establishment of best practices and standards for generative AI development within the organization.

Providing technical mentorship and guidance to junior team members.

Key Requirements of a Generative AI Engineer

5+ years of Industry experience primarily related to Unstructured Text Data and NLP (PhD work and internships will be considered if they are related to unstructured text in lieu of industry experience but not more than 2 years will be accounted towards industry experience) and 6 months+ Generative AI experience.

Generative AI certification, course work and evidence of relevant GenAI/Deep Learning courses available in the market (courses completion from any renowned LLM companies NVIDIA, Cohere, Predibase, Pinecone, Microsoft, Google, databricks)

Understanding of NLP concepts like sequence tagging, POS, sentiment analysis, machine translation, summarization

You have solid understanding of deep learning architecture like encoder-decoder (T5), encoder only transformers (BERT family of models), decoder only transformers, recurrent neural network, seq2seq models, LSTM, GPT, VAE, and GANs.

You are proficient in understanding Transformers, Word Embedding, Positional Encoding, Attention models and strong knowledge of maths/algebra to understand deep knowledge on latest transformers technique.

You are proficient in Python and have experience with machine learning libraries and frameworks such as TensorFlow, PyTorch, or Keras.

You have strong knowledge of data structures, algorithms, and software engineering principles.

You are familiar with cloud-based platforms and services, such as AWS, GCP, or Azure.

You have experience with natural language processing (NLP) techniques and tools, such as SpaCy, NLTK, or Hugging Face.

You have knowledge of software development methodologies, such as Agile or Scrum.

You possess excellent problem-solving skills, with the ability to think critically and creatively to develop innovative AI solutions.

You have strong communication skills, with the ability to effectively convey complex technical concepts to a diverse audience.

You possess a proactive mindset, with the ability to work independently and collaboratively in a fast-paced, dynamic environment.

Experience with developing and deploying products in production with experience in two or more of the following languages (Python, C++, Java, Scala)

Academic Qualifications:

Master\xe2\x80\x99s degree or above in Computer Science, Computational Linguistics, Mathematics, Physics, or Electrical Engineering with research experience from a strong academic program along with thesis

Completion of thesis/research is required as part of graduation in computer science, artificial intelligence, Mathematics, Physics, Electrical Engineering or statistics.

Nice to Have

A PhD degree in Computer Science, Artificial Intelligence, Computational Linguistics, Machine Learning, or related technical field from a strong academic program

Medical concepts with codes from standard ontologies (SNOMED CT, LOINC, RxNorm, ICD, etc.)

Experience with Kubernetes and dockers

Experience building REST API\xe2\x80\x99s for AI work and knowledge of microservices architecture

Participation in open source community projects

Publication record in top NLP conferences (NIPS, ICLR, ACL, NAACL, EMNLP, SIGIR, WWW, etc.)

Mandatory Requirements:

We require that all employees have the following technical capability at their home: High speed internet over 10 Mbps, the ability to plug in directly to the home internet router. These at-home technical requirements are subject to change with any scheduled re-opening of our office locations.

\xe2\x80\x8b

Evolent Health is an equal opportunity employer and considers all qualified applicants equally without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, veteran status, or disability status.

Evolent Health