Intern AI Engineer

About us:

For over 20 years, Smart Data Solutions has been partnering with leading payer organizations to provide automation and technology solutions enabling data standardization and workflow automation. The company brings a comprehensive set of turn-key services to handle all claims and claims-related information regardless of format (paper, fax, electronic), digitizing and normalizing for seamless use by payer clients. Solutions include intelligent data capture, conversion and digitization, mailroom management, comprehensive clearinghouse services and proprietary workflow offerings. SDS’ headquarters are just outside of St. Paul, MN and leverages dedicated onshore and offshore resources as part of its service delivery model. The company counts over 420 healthcare organizations as clients, including multiple Blue Cross Blue Shield state plans, large regional health plans and leading independent TPAs, handling over 500 million transactions of varying types annually with a 98% customer retention rate. SDS has also invested meaningfully in automation and machine learning capabilities across its tech-enabled processes to drive scalability and greater internal operating efficiency while also improving client results.  

SDS recently partnered with a leading growth-oriented investment firm, Parthenon Capital, to further accelerate expansion and product innovation.
 

Location: 6th Floor, Block 4A, Millenia Business Park, Phase II MGR Salai, Kandanchavadi, Perungudi, Chennai- 600096

Smart Data Solutions is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to race, color, sex, sexual orientation, gender identity, religion, national origin, disability, veteran status, age, marital status, pregnancy, genetic information, or other legally protected status. To perform this job successfully, an individual must be able to perform each essential duty satisfactorily. The requirements listed above are representative of the knowledge skill and or ability required. Reasonable accommodation may be made to enable individuals with disabilities to perform essential job functions.  Due to access to Protected Healthcare Information, employees in this role must be free of felony convictions on a background check report.  

Duties and Responsibilities include but are not limited to:

• Develop and evaluate machine learning and deep learning models for pharma-related document analysis.
• Build pipelines for OCR/OMR and extract structured information from unstructured text (clinical forms, prescriptions, regulatory filings, etc.).
• Integrate and experiment with LLMs like Qwen, Nuextract, and other open-source foundation models using LangChain.
• Write scalable, testable code in Python and occasionally in Java, especially for back-end or integration tasks.
• Assist in creating prompt templates, retrieval-augmented generation (RAG), and LLM-enhanced search capabilities.
• Support data cleaning, annotation, and labeling tasks in medical/NLP datasets.
• Collaborate with data scientists and domain experts to improve model accuracy and performance.

The duties set forth above are essential job functions for the role. Reasonable accommodations may be made to enable individuals with disabilities to perform essential job functions.

Skills and Qualifications

• Strong coding proficiency in Python and Java.
• Solid understanding of machine learning, deep learning, and NLP fundamentals.
• Exposure to or interest in LLMs, LangChain, Qwen, Nuextract, or other instruction-following/foundation models.
• Hands-on experience or coursework in OCR/OMR, computer vision, and document data extraction.
• Familiarity with libraries such as Transformers (Hugging Face), OpenCV, Tesseract, SpaCy, or PyTorch/TensorFlow.
• Ability to work independently and collaborate in cross-functional teams.
• Background in Biomedical AI, Healthcare Informatics, or Pharmaceutical NLP projects.
• Experience working with large PDF/TIFF document corpora and associated annotation tools.
• Knowledge of information retrieval, prompt engineering, and LLM deployment techniques.