Vetted Vector Search Professionals

“AI engineer with experience taking a GPT-4-powered GenAI career coach toward production on Azure AI Foundry, re-architecting the backend with hybrid (vector + keyword) search and RAG optimizations to cut latency by 50%. Also has client-facing TCS experience building healthcare ETL pipelines and delivering error-free monthly reports, plus current work analyzing agentic system reasoning traces and guardrail drift as an AI research fellow.”

“Data/ML engineer in financial services (Northern Trust) who built a production RAG-based LLM system to connect structured transaction/portfolio data with unstructured market and internal documents for risk teams. Strong in end-to-end pipelines (AWS Glue/Airflow/PySpark), entity resolution, and taking models from prototype to reliable daily production with performance tuning (LoRA + TensorRT) and monitoring.”

“Built and shipped a production RAG system (“TROJAN KNOWLEDGE”) for answering questions over technical PDFs, using a 3-stage retrieval stack (BM25 + FAISS + cross-encoder) to lift F1 from 71% to 84%. Drove major performance gains with a 3-level cache (memory/Redis/disk) cutting latency from ~200ms to ~10ms, and added Prometheus/Grafana monitoring plus LangChain-based fallback logic to handle OpenAI rate limits under load.”

“Built and deployed a production RAG pipeline at PNC Financial Services to let risk/compliance analysts query millions of internal financial documents in natural language, reducing manual search and speeding regulatory validation. Demonstrates deep practical experience with large-scale document ingestion/OCR cleanup, retrieval performance tuning (hierarchical indexing, caching), and LLM reliability controls (grounding, citations, abstention), plus cloud orchestration on Azure and AWS.”

“LLM/agent production specialist with strong customer-facing and pre-sales chops: turns demo-grade prototypes into reliable, compliant deployments using RAG tuning, guardrails, evals in CI, and observability with staged rollouts/rollback. Known for engineering-first workshops (including live break-and-fix on retrieval misses, tool timeouts, and prompt injection) that win over skeptical senior developers and drive adoption.”

“Internship experience shipping production AI systems: built an end-to-end RAG platform (Python/FastAPI + LangChain/LangGraph + vector search) to answer support questions from unstructured internal docs, with a strong focus on hallucination prevention through confidence gating and rigorous offline/online evaluation. Also delivered an AI-driven personalization/analytics feature using an unsupervised clustering pipeline, iterating with PMs to align statistically strong clusters with actionable business segmentation.”

“Built and deployed a production Task Prioritization App using Python/Streamlit/MongoDB with Gemini API to score and rank tasks by context (deadlines, dependencies, urgency). Focused on reliability challenges like prompt tuning for nuanced task understanding, concurrent DB updates, and performance via async LLM calls, and validated usability through iterative feedback with a non-technical end user.”

“Built and productionized an agentic LLM automation system for an insurance client to determine medication eligibility, using prompt-chaining plus a RAG pipeline over policy rules and deploying on AWS (Lambda/Step Functions, Bedrock) with a serverless architecture. Addressed major data/schema mismatch issues via a semantic matching pipeline and validated performance through human agreement scoring, A/B testing, KPI monitoring, and confidence-based human-in-the-loop review.”

“AI/ML engineer who has shipped production LLM and ML systems, including a RAG pipeline that ingested ~500k insurance/client documents to help adjusters answer questions faster and more consistently. Experienced in handling messy real-world document formats, tuning retrieval/chunking, and reducing latency via vector search optimization, precomputed embeddings, and caching. Also built orchestrated fraud-detection deployment workflows using AWS Step Functions and SageMaker, and partners closely with non-technical operations teams on NLP automation.”

“Built and deployed a RAG-based student/faculty support chatbot at a university that answers from official syllabus/policy documents and now supports 4,000+ students while reducing repetitive support requests. Hands-on with LangChain, LangGraph, and CrewAI to orchestrate reliable agentic workflows, with a strong focus on testing/monitoring in production and cross-functional delivery (e.g., marketing analytics automation at Steve Madden).”

“Built and deployed a production LLM-powered phone assistant for a healthcare clinic, combining streaming STT/TTS with RAG over approved clinic documents and strict safety guardrails to prevent unverified medical advice, plus seamless human handoff. Also has hands-on Apache Airflow experience building robust daily ML/data pipelines with data validation, retries/timeouts, monitoring, and metric-gated model deployment, and iterates closely with clinic staff using real call reviews.”

“Built a production, real-time insurance claims document-understanding and fraud-detection pipeline using TensorFlow + fine-tuned BERT, deployed on AWS (SageMaker/Lambda/API Gateway) with automated retraining via MLflow and Jenkins. Addressed noisy documents and latency using augmentation and model distillation (3x faster), cutting claims ops manual review by ~50% and reducing fraudulent payouts.”

“Built and deployed an AI agent to help patients navigate complex housing information by scraping and normalizing unstructured data across all 50 U.S. states, then layering a LangChain RAG system with MMR re-ranking to reduce hallucinations. Experienced in orchestrating multi-agent workflows (LangGraph/CrewAI) and production reliability practices (Pydantic-validated outputs, LLM-as-judge evals, tracing). Also delivered stakeholder-facing explainability via SHAP dashboards for a loan-approval predictive model at Welspot.”

“ML engineer/data scientist who deployed a production credit risk + insurance claims triage platform at Hartford Financial, combining XGBoost default prediction with BERT-based document classification. Demonstrated strong MLOps by cutting inference latency to sub-500ms and building drift monitoring plus automated retraining/deployment pipelines (MLflow, CloudWatch, GitHub Actions, SageMaker) with human-in-the-loop review and SHAP-based explainability for underwriting adoption.”

“Built and deployed NeuroDoc, a production-grade RAG system for PDF Q&A that delivers citation-backed answers with strong anti-hallucination guardrails. Experienced in orchestrating and scaling ML/LLM pipelines with Kubernetes, Airflow/Prefect, and PyTorch Distributed, and in building rigorous evaluation and citation-verification tooling to ensure reliability in production.”

“Built a real-time telemedicine clinician dashboard and iterated post-launch by diagnosing lag via logs/metrics and optimizing DB queries/sync logic. Also shipped a production internal RAG knowledge assistant for support teams, including embeddings/vector DB, citation-only answers with abstention thresholds, and an eval loop driven by real ticket data that improved accuracy through chunking/overlap and batching optimizations.”

“Backend/data engineer who built a production-ready multi-agent financial intelligence system (Mycroft) that orchestrates specialized AI agents to analyze real-time market data using FastAPI and Pinecone vector search. Brings strong security/reliability instincts (rate limiting, JWT/OAuth2, retries/backoff, health checks) and has caught high-impact data integrity issues in financial migrations (timezone normalization across global legacy systems).”

“Backend/ML platform engineer who built an end-to-end news summarization and personalized recommendation system using FastAPI, Redis, and a vector search pipeline (FAISS). Strong in productionizing services on Kubernetes with GitOps (ArgoCD + GitHub Actions), including CI image tagging/publishing and safe rollouts, plus experience migrating EC2 services to containerized orchestration with robust health checks and latency/error monitoring.”