What are the responsibilities and job description for the AI/ML Engineer position at Resource Informatics Group, Inc?
Role: AI/Client Engineer
Location: Hartford, CT 06103 (Day 1 Onsite) - Only Local candidate
Duration: Long Term contract
Skill Matrix to be filled by Candidates:
Mandatory SkillsYears of ExperienceYear Last UsedRating Out of 10Google Vertex AI Python Coding NLP GCP vector indexing LLM
Location: Hartford, CT 06103 (Day 1 Onsite) - Only Local candidate
Duration: Long Term contract
Skill Matrix to be filled by Candidates:
Mandatory SkillsYears of ExperienceYear Last UsedRating Out of 10Google Vertex AI Python Coding NLP GCP vector indexing LLM
- 4-5 Years of AI/Client experience.
- Python: Expertise in Python Data Exploration and Data Science stack - Jupyter Notebook, Pandas, Matplotlib, Sci-kit Learn etc.
- NLP: Experience using Hugging Face pipelines to perform various NLP tasks such as classification, generation, entity detection, etc.
- LLM Application: Hands-on experience using Llama Index or Lang chain to build semantic search, retrieval augmented generation (RAG), hybrid search systems.
- Prompt Engineering: Experience using Open AI or Vertex AI or Llama APIs to design and structure the inputs to an LLM programmatically.
- Vector Database: Experience using Vector Databases such as PineCone, Qdrant, Vespa, Weaviate, etc.
- Evaluation: Familiarity with NLP evaluation metrics used to assess retrieval and generation quality
- Cloud: Experience using Big cloud providers such as AWS, GCP, Azure to quickly deploy POCs.
- Familiarity with MongoDB Atlas data modeling, indexing, and querying.
- Familiarity with conversation AI platforms such as Kore AI, RASA, Google Dialog flow, CCAI, etc
- Experience using Approximate Nearest Neighbor libraries such as FAISS, ANNOY, etc.
- Familiarity with advanced Prompting techniques such as Few-shot learning, Chain-of-thought, etc. and leverage various features such as function calling, Responsible AI, etc.
- Familiarity with improving the vector indexing, Query Expansion, Cross-encoder reranking, Training and utilizing Embedding Adapters.
