- Senior
- Oficina en Pune
You will also be expected to apply modern LLMOps practices, handle schema-constrained generation, optimize cost and latency trade-offs, mitigate hallucinations, and ensure robust safety, personalization, and observability across GenAI systems.
- Design and implement scalable and modular pipelines for data ingestion, transformation, and orchestration across GenAI workloads.
- Manage data and model flow across LLMs, embedding services, vector stores, SQL sources, and APIs.
- Build CI/CD pipelines with integrated prompt regression testing and version control.
- Use orchestration frameworks like LangChain or LangGraph for tool routing and multi-hop workflows.
- Monitor system performance using tools like Langfuse or Prometheus.
- Data and Document Ingestion
- Develop systems to ingest unstructured (PDF, OCR) and structured (SQL, APIs) data.
- Apply preprocessing pipelines for text, images, and code.
- Ensure data integrity, format consistency, and security across sources.
- AI Service Integration
- Integrate external and internal LLM APIs (OpenAI, Claude, Mistral, Qwen, etc.).
- Build internal APIs for smooth backend-AI communication.
- Optimize performance through fallback routing to classical or smaller models based on latency or cost budgets.
- Use schema-constrained prompting and output filters to suppress hallucinations and maintain factual accuracy.
- Build hybrid RAG pipelines using vector similarity (FAISS/Qdrant) and structured data (SQL/API).
- Design custom retrieval strategies for multi-modal or multi-source documents.
- Apply post-retrieval ranking using DPO or feedback-based techniques.
- Improve contextual relevance through re-ranking, chunk merging, and scoring logic.
- LLM Integration and Optimization
- Manage prompt engineering, model interaction, and tuning workflows.
- Implement LLMOps best practices: prompt versioning, output validation, caching (KV store), and fallback design.
- Optimize generation using temperature tuning, token limits, and speculative decoding.
- Integrate observability and cost-monitoring into LLM workflows.
- Backend Services Ownership
- Design and maintain scalable backend services supporting GenAI applications.
- Implement monitoring, logging, and performance tracing.
- Build RBAC (Role-Based Access Control) and multi-tenant personalization.
- Support containerization (Docker, Kubernetes) and autoscaling infrastructure for production.
- Bachelor’s or Master’s in Computer Science, Artificial Intelligence, Machine Learning, or related field.
- 5+ years of experience in AI/ML engineering with end-to-end pipeline development.
- Hands-on experience building and deploying LLM/RAG systems in production.
- Strong experience with public cloud platforms (AWS, Azure, or GCP).
- Proficient in Python and libraries such as Transformers, SentenceTransformers, PyTorch.
- Deep understanding of GenAI infrastructure, LLM APIs, and toolchains like LangChain/LangGraph.
- Experience with RESTful API development and version control using Git.
- Knowledge of vector DBs (Qdrant, FAISS, Weaviate) and similarity-based retrieval.
- Familiarity with Docker, Kubernetes, and scalable microservice design.
- Experience with observability tools like Prometheus, Grafana, or Langfuse.
- Knowledge of LLMs, VAEs, Diffusion Models, GANs.
- Experience building structured + unstructured RAG pipelines.
- Prompt engineering with safety controls, schema enforcement, and hallucination mitigation.
- Experience with prompt testing, caching strategies, output filtering, and fallback logic.
- Familiarity with DPO, RLHF, or other feedback-based fine-tuning methods.
- Strong analytical, problem-solving, and debugging skills.
- Excellent collaboration with cross-functional teams: product, QA, and DevOps.
- Ability to work in fast-paced, agile environments and deliver production-grade solutions.
- Clear communication and strong documentation practices.
- Hands-on experience with basic NLP tasks such as Named Entity Recognition (NER), text classification, summarization, and entity linking.
- Practical exposure to traditional Machine Learning algorithms such as Logistic Regression, Random Forest, XGBoost, K-Means, DBSCAN, etc.
- Experience implementing and fine-tuning Deep Learning models (e.g., CNNs, RNNs, LSTM, Transformers) using frameworks like PyTorch or TensorFlow.
- Familiarity with OCR, document parsing, and layout-aware chunking techniques.
- Hands-on with MLOps and LLMOps tools for Generative AI
- Contributions to open-source GenAI or AI infrastructure projects.
- Knowledge of GenAI governance, ethical deployment, and usage controls.
- Experience with hallucination suppression frameworks like Guardrails.ai, Rebuff, or Constitutional AI.