Suyog S.

WORK HISTORY

EDUCATION

ABOUT SUYOG S.

I build production-grade Generative AI systems powered by large language models (LLMs), retrieval architectures, and multi-agent workflows.Currently leading AI/ML initiatives in financial services, I design and deploy scalable LLM platforms — with governance, reliability, and performance built in.🧠 Generative AI & LLM Systems• Retrieval-Augmented Generation (RAG) with hybrid search and dense vector indexing (FAISS, Snowflake vector capabilities)• Multi-agent orchestration using LangChain, LangGraph, and structured workflow design• Transformer architectures (BERT, T5, GPT), embeddings, fine-tuning & post-training optimization• Prompt engineering frameworks, evaluation pipelines, hallucination mitigation, and guardrail design• LLM evaluation and monitoring: bias detection, explainability (SHAP, LIME), fairness audits (Fairlearn) Production ML & Infrastructure• Real-time inference systems with latency optimization and scalable API design (FastAPI, Streamlit)• Vertex AI, Snowflake, BigQuery, Hugging Face Transformers• MLOps: CI/CD for ML, model versioning, drift detection, monitoring, secure deployment• Enterprise controls: PII masking, access governance, audit logging, model risk documentation What I Focus OnI operate at the intersection of engineering, product, and risk — translating ambiguous business problems into robust AI architectures that deliver measurable impact.My recent work includes:• Building agentic AI systems for natural language-to-SQL analyticsI am particularly interested in roles where I can:• Architect LLM platforms at scale• Advance production RAG and multi-agent systems• Contribute to Responsible AI and evaluation frameworks• Partner deeply with engineering teams to ship real-world AI systemsBuilding in public: github.com/suyog957