Solutions Architect, Bengaluru

The Solution Architect will be responsible for designing and implementing end-to-end AI architectures that integrate Large Language Models (LLMs) into production-ready products. This role sits at the intersection of software engineering, data science, and infrastructure, focusing on creating scalable, secure, and cost-effective AI solutions.Key Responsibilities: Architecture Design: Lead the design of RAG (Retrieval-Augmented Generation) architectures, agentic workflows, and multi-model systems. Product Customization: Translate business requirements into technical blueprints for fine-tuning models or utilizing prompt engineering to meet specific product goals.Evaluation&Optimization: Establish frameworks for evaluating model performance (e.g., faithfulness, relevancy) and optimize for inference latency and token costs. Infrastructure&Integration: Design the integration between LLMs, vector databases (e.g., Pinecone, Milvus, or Weaviate), and existing enterprise data pipelines. Security&Compliance: Ensure all AI implementations adhere to data privacy standards, focusing on PII masking and preventing prompt injection or data leakage. Technical Leadership: Act as the bridge between the Product Owner and the Engineering team, ensuring technical feasibility and long-term maintainabilityRequired Technical Skills: Bachelors or Masters in Computer Science with 10 years of experience in the field&atleast 3+ years in technical architecture design for AI solutions LLM Frameworks: Deep proficiency in LangChain, LlamaIndex, or Haystack . Model Providers: Experience working with APIs from OpenAI, Anthropic, or Google (Gemini/Vertex AI), as well as open-source models (Llama 3, Mistral). Vector Databases: Practical experience with vector embeddings and similarity search.Cloud&DevOps: Strong knowledge of cloud AI platforms (AWS Bedrock, Azure AI Studio, or Google Vertex AI) and CI/CD for ML (MLOps). Coding: Proficiency in Python (fastAPI, Pydantic) and understanding of asynchronous programming Preferred Qualifications: Experience in Fine-tuning (PEFT/LoRA) and quantization techniques.Background in traditional NLP (Named Entity Recognition, Sentiment Analysis). Familiarity with guardrail frameworks (e.g., NeMo Guardrails or Guardrails AI). Job Locations Available: Mumbai, Bangalore, Pune, Hyderabad, Gurgaon