AI/ML Engineer

FutureNexas Labs (Gen AI & ML Engineer), Canada                                May 2025 – Present

AI-Powered Chatbot, Matching Algorithm, and Intelligent Broadcasting System

Objective: Developed LLM-driven chat and ML systems to automate communication, optimize matching between users, and improve broadcast relevance in hyper-local applications.

Key Contributions:

• Built a production-grade LLM Chatbot (RAG + Gemini API) that automates onboarding, FAQs, and support flows by retrieving context from internal knowledge bases using embeddings, vector search, and LangChain.

• Designed a multi-stage Text Moderation Pipeline combining regex filters, lightweight classifiers, and transformer models to detect toxicity, hate, threats, and inappropriate content with significantly improved precision/recall.

• Engineered a Smart Broadcast Ranking System that scores and prioritizes helpers based on geolocation, detour distance, trust score, past behavioral patterns, and demand signals improving match success rate and notification relevance.

• Developed a Real-Time Matching Algorithm using OSRM routing, proximity scoring, radius expansion logic, helper/request compatibility scoring, and Redis caching for sub-second match computation.

• Implemented MLOps Pipelines with MLflow, Airflow, and Vertex AI for automated feature generation, retraining, model versioning, and continuous evaluation.

• Containerized and Deployed microservices on GCP Cloud Run with Docker, CI/CD pipelines, Cloud Build triggers, and integrated monitoring via Cloud Logging and Prometheus.

• Collaborated with backend engineers to define API contracts for matching, broadcast, chatbot, and moderation services while ensuring scalable, fault-tolerant architecture.

Data Analyst

PNC Bank (Finance | Credit Risk Analytics), USA, Remote                  Aug 2022 - May 2023
Credit Risk Modeling & Customer Segmentation

Objective: Improve credit risk scoring by building predictive models and dashboards for underwriting and customer segmentation.

Key contributions:

• Collected and processed customer demographic, credit history, and transaction data.

• Used SQL and Python to clean, normalize, and handle missing or inconsistent records.

• Performed univariate and multivariate analysis to detect trends in payment patterns, default triggers, and high-risk behaviors.

• Visualized data distributions and segment overlaps using boxplots, heatmaps, and feature importance graphs.

• Created features from payment delays, balance utilization, loan-to-income ratio, and frequency of missed payments.

• Reduced dimensionality using PCA and removed collinear variables via correlation matrix analysis.

• Trained and tuned models including Logistic Regression, Random Forest, and XGBoost to predict default probability and rank customer risk.

• Achieved high AUC (~0.85) while reducing false positives by 20%.

• Helped underwriters identify risky profiles early and prioritize customer outreach.

• Enabled more personalized insurance offerings through customer clustering and segmentation.

• Orchestrated data ingestion, model training, and scoring using Airflow, with real-time scoring in AWS Lambda.

• Integrated performance monitoring with AWS CloudWatch, setting up alerts for drift and data anomalies.

• Developed Power BI dashboards that visualized credit risk scores, prediction confidence, and customer clusters.

AWS DevOps Engineer

Techolution India Pvt Ltd, Hyderabad, India                                           Dec 2019 – Aug 2022

Client: Sonic Drive-In (USA)

Role: Cloud Platform Support and Observability Engineer

Started as DevOps Intern; transitioned into a full-time AWS Engineer role after consistently delivering automated infrastructure improvements.

Key Contributions:

• Built scalable, secure, and compliant cloud infrastructure across AWS accounts using Terraform, Ansible, and AWS CloudFormation.

• Designed and deployed CI/CD pipelines with Jenkins, GitHub Actions, GoCD, and Bitbucket, debugging issues across the full pipeline lifecycle.

• Containerized applications with Docker and managed production-ready Kubernetes workloads on EKS, ensuring high availability and auto-scaling.

• Managed ECS and EKS containers for production and testing environments.

• Automated infrastructure provisioning and deployment workflows via AWS CodePipeline and AWS Lambda.

• Maintained Agile documentation in Confluence and managed delivery workflows via JIRA.

• Handled L2 incident management using ServiceNow and implemented end-to-end tracking.

• Used Auth0 for secure authentication and customer identity management.

• Deployed Android and iOS updates and monitored releases through Splunk and Dynatrace to capture performance and crash analytics.

• Deployed ML inference services and microservices using AWS ECS and Fargate, implementing blue-green deployments and rollbacks.

• Monitored system health with CloudWatch, ELK Stack, Prometheus, and Grafana; set up custom alarms for latency and throughput SLIs.

• Used IAM roles and KMS to secure APIs, Lambda functions, and S3 access patterns. Audited and enforced least-privilege access models.

• Actively participated in on-call rotations, supported platform migrations, and led RCA documentation for incident reviews.