Available for Staff+ / Principal roles · open to consulting
Seattle, WA · 47.6°N 122.3°W

Vybhav Reddy KC.

Staff Data Scientist building agentic AI systems for identity verification, fraud detection, and risk — at production scale, in regulated environments.

10+
Years
$5.35M
Quarterly Rev.
500K+
Daily Verifications
Book a 15-min chat Download Resume LinkedIn GitHub
GENAIAGENTIC AIFRAUD DETECTIONIDENTITY VERIFICATIONFINTECHNLPMLOPSRAGLLM EVALSPRODUCT SENSE GENAIAGENTIC AIFRAUD DETECTIONIDENTITY VERIFICATIONFINTECHNLPMLOPSRAGLLM EVALSPRODUCT SENSE
01 — About

A decade at the seam between research and production.

I lead data science work at Socure, where I design and ship machine-learning systems that verify identities and stop fraud for top US banks and fintechs.

My recent work has focused on agentic AI — orchestrating multi-agent workflows for autonomous decisioning, continuous-learning pipelines, and explainable risk models that meet FCRA and GDPR requirements.

Before Socure I spent five years at Conduent shipping NLP, computer-vision, and document-intelligence products across legal, healthcare, and ecommerce.

— Scope of role Beyond the modeling work: I mentor junior data scientists, lead technical strategy for the identity org, partner cross-functionally with product, engineering, sales, and legal on FCRA / GDPR compliance, and present roadmap and revenue impact to executive stakeholders.
02 — Experience

Where I've built things that mattered.

03 — Selected Work

Production systems with measured impact.

/ 01

Agentic AI for Identity Verification

Autonomous multi-agent workflow orchestrating data validation, decisioning, and continuous learning. Hybrid rule-based + ML + agent reasoning for accuracy and explainability.

+4% Gen-Z auto-verify <100ms p95 99.95% uptime
PYTHON · LANGGRAPH · GNN · KAFKA · FEATURE STORE
/ 02

LLM-Powered Reason Code Explainer

GenAI system translating SHAP-based feature importance into FCRA-compliant adverse-action notices. Multi-language, regulator-grade audit trails.

+15% annual revenue −40% review time +25% appeal success
LLMS · SHAP · LIME · FCRA · RAG
/ 03

Watchlist & Sanctions Screening

Fuzzy-matching agent with adaptive thresholds, NLP-driven name-variation expansion, and phonetic algorithms for high-recall sanctions screening.

FPR 18% → <8% 99.9% recall
NLP · PHONETIC MATCHING · ELASTICSEARCH
/ 04

LLM-Based QA Evaluation System

Automated test-case generation and execution for ML products. Replaces manual QA pipelines with agent-driven evaluation, faithfulness scoring, and regression detection.

−20% manual QA +15% product quality
LLM EVALS · PYTHON · CI/CD
/ 05

Intelligent Case Management

Multi-agent intake, routing, resolution, and QA pipeline using sentence-transformer embeddings for case similarity and priority-queue optimization.

35% E2E automation 4.2h → 1.8h resolution FCR 62% → 84%
AGENTS · SENTENCE-TRANSFORMERS · NLP
/ 06

Financial Contracts Analyzer

ML/AI pipeline parsing hundreds of pages of contracts to surface monetary risk and savings opportunities. Action-oriented insights for legal and finance teams.

~$1.2M savings
NLP · DOCUMENT AI · TRANSFORMERS
04 — Voices

What people I've worked with say.

"
Vybhav consistently demonstrated the rare ability to translate complex ML research into production systems that moved the business — his work on identity verification directly impacted top-tier bank clients and the broader fraud roadmap.
Engineering Leader Socure · Identity & Fraud
"
He doesn't just build models — he frames the problem, aligns the org, and drives outcomes. The agentic AI work he led was genuinely ahead of the curve, and he made it accessible to product and legal teams alike.
Product Partner Cross-functional · Risk Products
"
One of the most thoughtful data scientists I've worked with. Strong technical depth across NLP, ML, and GenAI, paired with the patience to mentor and the clarity to present to executives.
Senior Data Scientist Former Conduent colleague
"
Vybhav owned the analytics strategy end-to-end — from the dashboard that lifted KPIs by 35% to the pre-sales custom models that drove millions in pipeline. Rare combination of rigor and commercial sense.
Senior Manager Socure · Customer Strategy

— More on LinkedIn recommendations ↗

— Featured in / Worked on
05 — Stack

Tools, languages, and territory.

Areas of Expertise

  • GenAI & Agentic AI
  • Machine Learning
  • Deep Learning
  • Natural Language Processing
  • Statistical Inference
  • Product Sense
  • Data Visualization

Languages & Frameworks

  • Python
  • SQL
  • PySpark
  • Scala
  • PyTorch
  • TensorFlow
  • XGBoost · LightGBM

Platform & Cloud

  • AWS · GCP · Azure
  • Databricks
  • Hadoop
  • Airflow
  • Salesforce
  • PowerBI · Looker
  • Kafka · Flink
— Let's talk

Building something
worth building?

vybhav19@gmail.com
Book a 15-min chat Download Resume