Bal mukund Jha

Architected and deployed an enterprise-grade Retrieval-Augmented Generation (RAG) system using LangChain and Pinecone, reducing hallucination rates by 40% and improving response accuracy for 50k+ daily active users. Designed scalable LLMOps pipelines with MLflow and Kubernetes, streamlining model fine-tuning and deployment cycles from weeks to days for a suite of domain-specific language models. Engineered a multi-agent orchestration framework utilizing CrewAI and AutoGPT, automating complex cross-departmental workflows and increasing operational efficiency by 35% across the organization. Led the migration of legacy on-premise ML workloads to AWS Bedrock, resulting in a 25% reduction in cloud compute costs while maintaining 99.9% system uptime during peak traffic. Implemented AI guardrails using NeMo Guardrails and custom prompt-injection filters, achieving 100% compliance with GDPR and enterprise data privacy policies for consumer-facing interfaces. Optimized model inference latency by 60% through the implementation of quantization (INT8) and NVIDIA TensorRT acceleration, supporting real-time decision-making for high-frequency trading applications. Developed unified vector database strategy (Weaviate/Milvus) for multi-modal data ingestion (text, image, audio), enabling 3x faster semantic search across internal knowledge bases. Collaborated with cross-functional stakeholders to define AI roadmaps and KPIs, successfully launching 4 mission-critical AI products that generated an estimated $12M in new annual recurring revenue. Spearheaded the integration of Federated Learning protocols for healthcare clients, allowing model training on sensitive datasets while ensuring zero-data leakage and strict HIPAA compliance. Mentored team of 15+ ML engineers and data scientists on system design best practices, cultivating AI-native engineering culture and increasing team velocity by 20% within 12 months.

Bachelor of Technology (B.Tech) in Mechanical Engineering DCRUST University, Murthal Sonepat, Harayana Expected December, 2026 • Completed 130/134 credits; pending final clearance of 1 subject Engineering Mathematics 1.