Senior ML Engineer Fraud, NLP & LLM Systems

Manish Vemula

Machine Learning Engineer

+1-734-***-**** ******.*****@*****.*** LinkedIn

Machine Learning Engineer with 5+ years of experience designing, deploying, and operating production-grade ML and LLM-powered systems in financial and enterprise environments. Specializes in NLP, fraud detection, Retrieval-Augmented Generation (RAG), and scalable inference services built on AWS and GCP. Strong track record of owning the full ML lifecycle—from problem framing and feature engineering to deployment, monitoring, and performance optimization—delivering measurable business impact through reliable and cost-efficient AI systems. TECHNICAL SKILLS

Machine Learning & AI: Supervised & Unsupervised Learning, Deep Learning, Transformers, LLM Integration, Retrieval-Augmented Generation

(RAG), NLP, Time Series Modeling, Anomaly Detection, Ensemble Methods, Distributed Training, Batch & Streaming Inference, GPU Acceleration LLM & NLP Stack: Hugging Face Transformers, LangChain, BERT, Prompt Engineering, Vector Embeddings, FAISS MLOps & Production ML: MLflow (Model Registry), CI/CD for ML, Drift Detection, Experiment Tracking, Model Monitoring, Docker, Kubernetes, Data Versioning (DVC), Feature Store Concepts, Data Validation (Great Expectations), Model Lineage Tracking Cloud & Infrastructure: AWS SageMaker, AWS Lambda, S3, EC2, GCP Vertex AI, BigQuery, Terraform, CloudFormation Data Engineering & APIs: Python, FastAPI, Flask, SQL (PostgreSQL, MySQL), Pandas, NumPy, ETL Pipelines, RESTful APIs, Apache Kafka Statistical Methods: A/B Testing, Hypothesis Testing, Regression, Classification, Clustering, Dimensionality Reduction EXPERIENCE

Machine Learning Engineer Discover Financial Services – Remote, USA Sept 2023 – Present Architected and productionized machine learning services supporting fraud detection and customer intelligence platforms within a cloud-native environment.

• Designed an end-to-end fraud detection framework leveraging LSTM networks and Variational Autoencoders, increasing anomaly detection precision by 30% and reducing false positives in high-volume transaction streams.

• Built a Retrieval-Augmented Generation (RAG) pipeline leveraging transformer embeddings, vector similarity search (FAISS), prompt optimization, and structured document chunking; integrated batch ingestion workflows and streaming inference endpoints to support scalable knowledge retrieval.

• Developed RESTful inference APIs using FastAPI with Dockerized deployment, enabling real-time and batch prediction workflows; optimized GPU utilization and autoscaling policies to maintain low-latency performance under variable traffic loads.

• Integrated MLflow model registry with data versioning practices (DVC) to ensure reproducible training pipelines, controlled promotion across environments, and auditable model lineage.

• Automated CI/CD pipelines via GitHub Actions and Jenkins to support model validation, container builds, and controlled rollouts, decreasing deployment time by 30%.

• Provisioned and optimized cloud infrastructure on AWS SageMaker and GCP Vertex AI, introducing autoscaling policies that lowered compute costs by 25% without performance degradation.

• Implemented data validation and drift detection pipelines using statistical distribution monitoring and schema validation (Great Expectations), triggering automated retraining workflows to maintain SLA-aligned accuracy.

• Designed a microservices-based ML architecture separating feature engineering, training, and inference layers; leveraged message-driven pipelines (Kafka) to support streaming data ingestion and fault-tolerant scaling. Associate Machine Learning Engineer Nexova – India May 2019 – July 2022 Delivered scalable ML and NLP systems across enterprise applications, progressing from model development to production deployment ownership.

• Led development of transformer-based NLP systems for chatbot sentiment analysis and intent classification, improving classification accuracy by 18% through optimized tokenization and contextual embeddings.

• Applied ensemble learning methods including XGBoost and Random Forest to forecasting and prediction systems, strengthening model robustness across structured datasets.

• Engineered automated ETL and feature engineering pipelines, introducing centralized feature storage patterns to improve feature reuse and consistency across multiple ML models.

• Deployed containerized ML services using Docker and AWS Lambda, enabling consistent cross-environment deployments and lowering infrastructure overhead.

• Optimized deep learning workflows using distributed training techniques and parallelized data loaders to improve training efficiency across large datasets.

• Migrated monolithic systems to microservices-based architecture, isolating ML inference services and improving uptime by 20% through fault-tolerant design.

• Implemented experiment tracking and performance benchmarking practices, enabling systematic evaluation of model improvements prior to release.

EDUCATION & CERTIFICATIONS

Master of Business Administration (MBA) Central Michigan University, MI (2024–2025) Master of Science in Information Systems Central Michigan University, MI (2022–2024) Bachelor of Computer Science JNTU Hyderabad, India

AWS Certified Machine Learning – Specialty

Microsoft Certified: Azure Data Scientist Associate

Contact this candidate