Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Hi, I’m Divya Sri, a Senior AI Engineer with 10+ years of experience building and deploying AI applications, LLM integrations, and GenAI systems across financial services, healthcare, telecom, and e-commerce. I design end-to-end GenAI solutions using GPT-4o, Claude 3.5 Sonnet, and Llama 3, delivering robust, governance-driven LLM-powered systems for high-stakes enterprise environments. I’ve architected production RAG pipelines with Pinecone, BM25 hybrid search, and cross-encoder reranking to boost retrieval accuracy by 28-40%. I’ve integrated OpenAI, Anthropic, and Cohere APIs, built structured output and JSON prompts, implemented LLM safety with Guardrails and NVIDIA NeMo, and established LLM Ops pipelines with LangSmith and LangGraph for monitoring, experimentation, and prompt regression testing. I’ve deployed AI across AWS, Azure, and GCP, enabling multi-geo, multi-tenant deployments, and consistently reduced production cycle times by 50-60% while increasing automation of high-volume document processing and decision-support workflows.…Hi, I’m Divya Sri, a Senior AI Engineer with 10+ years of experience building and deploying AI applications, LLM integrations, and GenAI systems across financial services, healthcare, telecom, and e-commerce. I design end-to-end GenAI solutions using GPT-4o, Claude 3.5 Sonnet, and Llama 3, delivering robust, governance-driven LLM-powered systems for high-stakes enterprise environments. I’ve architected production RAG pipelines with Pinecone, BM25 hybrid search, and cross-encoder reranking to boost retrieval accuracy by 28-40%. I’ve integrated OpenAI, Anthropic, and Cohere APIs, built structured output and JSON prompts, implemented LLM safety with Guardrails and NVIDIA NeMo, and established LLM Ops pipelines with LangSmith and LangGraph for monitoring, experimentation, and prompt regression testing. I’ve deployed AI across AWS, Azure, and GCP, enabling multi-geo, multi-tenant deployments, and consistently reduced production cycle times by 50-60% while increasing automation of high-volume document processing and decision-support workflows.

Divya Sri

AI Engineer, Data Scientist, AI Developer, +2





Hi, I’m Divya Sri, a Senior AI Engineer with 10+ years of experience building and deploying AI applications, LLM integrations, and GenAI systems across financial services, healthcare, telecom, and e-commerce. I design end-to-end GenAI solutions using GPT-4o, Claude 3.5 Sonnet, and Llama 3, delivering robust, governance-driven LLM-powered systems for high-stakes enterprise environments. I’ve architected production RAG pipelines with Pinecone, BM25 hybrid search, and cross-encoder reranking to boost retrieval accuracy by 28-40%. I’ve integrated OpenAI, Anthropic, and Cohere APIs, built structured output and JSON prompts, implemented LLM safety with Guardrails and NVIDIA NeMo, and established LLM Ops pipelines with LangSmith and LangGraph for monitoring, experimentation, and prompt regression testing. I’ve deployed AI across AWS, Azure, and GCP, enabling multi-geo, multi-tenant deployments, and consistently reduced production cycle times by 50-60% while increasing automation of high-volume document processing and decision-support workflows.…Hi, I’m Divya Sri, a Senior AI Engineer with 10+ years of experience building and deploying AI applications, LLM integrations, and GenAI systems across financial services, healthcare, telecom, and e-commerce. I design end-to-end GenAI solutions using GPT-4o, Claude 3.5 Sonnet, and Llama 3, delivering robust, governance-driven LLM-powered systems for high-stakes enterprise environments. I’ve architected production RAG pipelines with Pinecone, BM25 hybrid search, and cross-encoder reranking to boost retrieval accuracy by 28-40%. I’ve integrated OpenAI, Anthropic, and Cohere APIs, built structured output and JSON prompts, implemented LLM safety with Guardrails and NVIDIA NeMo, and established LLM Ops pipelines with LangSmith and LangGraph for monitoring, experimentation, and prompt regression testing. I’ve deployed AI across AWS, Azure, and GCP, enabling multi-geo, multi-tenant deployments, and consistently reduced production cycle times by 50-60% while increasing automation of high-volume document processing and decision-support workflows.

Available to hire

Hi, I’m Divya Sri, a Senior AI Engineer with 10+ years of experience building and deploying AI applications, LLM integrations, and GenAI systems across financial services, healthcare, telecom, and e-commerce. I design end-to-end GenAI solutions using GPT-4o, Claude 3.5 Sonnet, and Llama 3, delivering robust, governance-driven LLM-powered systems for high-stakes enterprise environments.

I’ve architected production RAG pipelines with Pinecone, BM25 hybrid search, and cross-encoder reranking to boost retrieval accuracy by 28-40%. I’ve integrated OpenAI, Anthropic, and Cohere APIs, built structured output and JSON prompts, implemented LLM safety with Guardrails and NVIDIA NeMo, and established LLM Ops pipelines with LangSmith and LangGraph for monitoring, experimentation, and prompt regression testing. I’ve deployed AI across AWS, Azure, and GCP, enabling multi-geo, multi-tenant deployments, and consistently reduced production cycle times by 50-60% while increasing automation of high-volume document processing and decision-support workflows.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

Work Experience

Senior AI/ML Engineer at JPMorgan Chase

August 1, 2024 - Present

Client engagement for JPMorgan Chase; evaluated financial research workflows across investment banking and credit risk, identifying AI opportunities and implementing a GenAI architecture for document reasoning, entity extraction, and risk intelligence. Benchmarked GPT-4o, Claude 3.5 Sonnet, and Gemini across financial QA and summarization tasks; established model selection criteria and prompt engineering baselines. Built Apache Kafka-based ingestion pipelines processing SEC filings, earnings transcripts, and Bloomberg feeds into a Delta Lake store, enabling retrievability and context enrichment for enterprise retrieval augmented generation. Developed production RA G pipelines with Pinecone, BM25 hybrid search, and Reranking to maximize retrieval precision; integrated with a structured JSON output for downstream analytics. Built prompt governance using LangSmith and Git versioning with regression testing; reduced prompt-related production incidents by 60-65%. Implemented end-to-end gove

Senior AI/ML Engineer at Liberty Mutual Insurance

July 1, 2022 - July 1, 2024

Led enterprise GenAI and LLM integration initiatives across financial services, health care, telecom, and e-commerce. Designed and productionized multi-tenant, geo-distributed LLM systems with guardrails, risk controls, and compliance considerations (HIPAA, FINRA). Built end-to-end RAG pipelines using Pinecone and BM25 hybrid search with cross-encoder reranking, improving retrieval precision by 28-40% over single-retrieval baselines. Implemented structured prompt frameworks, JSON mode outputs, and few-shot prompting to support reliable downstream systems. Integrated OpenAI, Anthropic, and Cohere APIs with robust orchestration; executed LLM fine-tuning (LoRA/QLoRA) on Llama 3 and Claude 3.5/4 to produce domain-specific models achieving improved benchmark results. Established LLM safety and governance using Guardrails, NVIDIA NeMo Guardrails, prompt injection detection, hallucination mitigation, and thorough output filtering at inference time. Built LLM Ops pipelines with LangSmith, Lang

AI/ML Engineer at Molina Healthcare

September 1, 2020 - July 1, 2022

Developed HIPAA-compliant NLP platform for clinical data; automated clinical document ingestion via HL7 FHIR APIs, standardizing CCD and CDA across 8 EHR integrations into a unified JSON schema. Built streaming pipelines ingesting 500K+ clinical documents using OpenAI embeddings, enabling semantic retrieval and reasoning. Implemented DAG-based pre-processing for data normalization, feature extraction, and de-identification. Created production RA G pipelines combining Pinecone, BM25 hybrid search, and RERANK to support retrieval-driven clinical decision support. Reduced clinician review time and accelerated routing decisions via RT dashboards and alerts; established governance with robust security and privacy controls.

ML Engineer at AT&T

September 1, 2018 - August 1, 2020

Assessed network operations workflows to identify and remediate delays in high-severity incidents. Built real-time ML anomaly detection across 50K+ devices using Spark-based pipelines, ingesting SNMP traps and performance logs for proactive fault isolation. Implemented streaming DAGs, online scoring, and alerting; reduced mean time to detect and resolve incidents, and minimized high-severity escalations. Established model retraining triggers, CI/CD pipelines, and monitoring dashboards to sustain production reliability.

Data Scientist / ML Engineer at Krios Info Solutions

July 1, 2015 - May 1, 2018

Evaluated e-commerce client’s legacy recommendation system; translated business requirements into ML problems and problem statements. Built PySpark ETL pipelines on Hadoop to ingest MySQL and Oracle data into HDFS for collaborative filtering model training. Engineered feature pipelines and a scalable ranking model using XGBoost, Word2Vec embeddings, and 500K+ customer interactions. Delivered a production NLP classifier and a 240+ category taxonomy for product tagging; achieved improvements in recommendation relevance and conversions.