About this role
Accenture seeks a full-time Gen AI Consultant in its Public Services Data & AI practice. The role is listed for Gurugram/Bengaluru and requires 5–10 years of experience. Travel may range from 0–100% according to client or project needs, including international client assignments.
Responsibilities
- Own and lead significant client workstreams, designing and developing generative AI solutions for document processing, customer interaction and automation, and integrating models into enterprise workflows.
- Run requirements workshops and feasibility assessments; translate business needs into technical specifications and fit-for-purpose Gen AI use cases.
- Design prompt-engineering frameworks, retrieval-augmented generation pipelines and metrics for accuracy and relevance. Work with cross-functional teams to integrate solutions with existing data ecosystems, APIs and cloud environments.
- Build reusable accelerators, templates and components; support client demonstrations, user testing and performance optimization; contribute to proposals and thought leadership.
Required expertise and capabilities:
- Deep understanding of LLM architectures, vector search and RAG pipelines; experience with OpenAI, Azure OpenAI, AWS Bedrock, Google Vertex AI or Anthropic Claude.
- Strong proficiency in Python, LangChain, prompt chaining, embedding models and API orchestration. Familiarity with enterprise integration, data pipelines and orchestration tools such as Airflow, Databricks and Azure Data Factory.
- Working knowledge of evaluation approaches including BLEU, ROUGE, accuracy and hallucination detection, plus model governance, bias mitigation and responsible AI practices.
- Ability to lead client outcomes, communicate through client-ready presentations, analyze complex problems and collaborate in multicultural teams. The practice serves public-sector areas including transport and infrastructure, defense, education, public health and safety, revenue and social services.
Skills for this role
Generative AILarge language modelsRetrieval-augmented generationVector searchPrompt engineeringPrompt chainingEmbedding modelsPythonLangChainAPI orchestrationOpenAIAzure OpenAIAWS BedrockGoogle Vertex AIAnthropic ClaudeAirflowDatabricksAzure Data FactoryData pipelinesEnterprise integrationModel evaluationBLEUROUGEHallucination detectionModel governanceBias mitigationResponsible AIRequirements gatheringClient presentationsLeadership