About this role
TransUnion CIBIL seeks a hands-on AI engineer for its Manager - Gen AI position in Mumbai. The role designs, builds and deploys AI solutions for operations and data science business problems, with an emphasis on implementation using modern LLM frameworks and integration with enterprise systems. It is a full-time, hybrid position requiring in-person work at an assigned TransUnion office at least two days per week.
Responsibilities
- Design and implement generative AI solutions using LLMs, retrieval-augmented generation (RAG) and agents for operations and data science use cases.
- Develop and deploy customer-focused AI/ML applications, including text classification and vector embeddings for structured, tabular data. Build scalable pipelines that bring structured and unstructured data into AI systems.
- Develop and optimize prompts, embeddings and retrieval systems. Partner with business teams to translate use cases into functional solutions, then integrate applications into existing workflows.
- Evaluate models for performance, cost, latency and accuracy, while applying governance, security and responsible AI practices.
Required qualifications
- A bachelor’s degree in engineering, statistics, economics, data science or a related field, plus 3–5 years of analytics experience. Operations, business or product analytics experience is preferred.
- Strong Python programming and hands-on experience with LLMs such as OpenAI, Azure OpenAI or open-source models including Llama and Mistral. Strong prompt engineering skills and understanding of TensorFlow and PyTorch, including model architecture, training, fine-tuning and optimization.
- Expertise in RAG architectures and vector databases such as Pinecone, FAISS or Weaviate; experience with LangChain, LlamaIndex or similar orchestration frameworks; and knowledge of machine learning fundamentals and NLP.
- Experience deploying on Azure, AWS or GCP, plus familiarity with APIs, microservices and data pipelines.
Skills for this role
Generative AILarge Language ModelsRetrieval-Augmented GenerationAI agentsPythonPrompt engineeringVector embeddingsVector databasesPineconeFAISSWeaviateLangChainLlamaIndexOpenAIAzure OpenAILlamaMistralTensorFlowPyTorchDeep learningMachine learningNatural language processingText classificationModel fine-tuningModel evaluationAzureAWSGoogle Cloud PlatformAPIsMicroservices ecosystems