Machine Learning Ml Engineer

Yeswanth Sai Yerramasu

Email: **************@*****.*** Phone: 443-***-****

PROFESSIONAL SUMMARY

AI/ML Engineer with 3+ years of experience delivering end-to-end machine learning solutions across healthcare and IT environments. Skilled in Python, SQL, R, and Java, with expertise in regression, classification, clustering, and advanced ML algorithms such as XGBoost and Random Forests. Proficient in deep learning frameworks (TensorFlow, PyTorch, Keras) and NLP/LLMs including BERT, RAG, and LangChain for clinical data processing. Experienced in building scalable data pipelines using PySpark, Airflow, dbt and deploying models via MLflow, Kubernetes, FastAPI, and Kubeflow. Strong cloud background spanning AWS SageMaker, GCP Vertex AI, and Azure ML, with hands-on experience in Snowflake, MongoDB, and ElasticSearch. Adept at designing compliant solutions aligned with HL7 FHIR, HIPAA, ICD-10, CPT, and PHI de-identification standards. Proven ability to integrate security and governance controls with model explainability frameworks (SHAP, LIME) to ensure transparency and regulatory alignment. SKILLS

• Languages: Python, SQL, R, Java, Bash

• Core ML: Regression, classification, clustering, XGBoost, Random Forests, model calibration

• Deep Learning: TensorFlow, PyTorch, Keras

• NLP and LLMs: Transformers, BERT, prompt engineering, RAG, LangChain

• Data & Pipelines: Pandas, NumPy, PySpark, Airflow, dbt, feature engineering

• MLOps: MLflow, Docker, Kubernetes, FastAPI, Git, CI/CD, Kubeflow, Flask, LLMOps, model monitoring

• Cloud: AWS SageMaker, S3, Lambda, GCP BigQuery and Vertex AI, Azure Machine Learning

• Datastores & Vector: PostgreSQL, Snowflake, MongoDB, ElasticSearch, FAISS, Pinecone

• Visualization: Tableau, Power BI, Plotly, Matplotlib

• Healthcare Standards & Compliance: HL7 FHIR, ICD-10, CPT, LOINC, HIPAA, PHI de-identification

• Security & Compliance: IAM, Data Governance, FedRAMP, SOC2 Alignment, Model Explainability (SHAP, LIME) EXPERIENCE

AI/ ML Engineer Globus Medical Inc Jan 2025 – Current

• Built clinical NLP pipelines with BERT to extract problems, meds, and procedures; mapped outputs to ICD-10/CPT/LOINC.

• Implemented PHI de-identification and audit trails; enforced HIPAA safeguards across data lakes and model endpoints.

• Developed computer-vision models (PyTorch/TensorFlow) for surgical workflow analytics, optimized inference latency by ~35%.

• Standardized FHIR resources (Patient, Encounter, Observation) to unify EHR feeds; reduced downstream data defects.

• Productionized models on Kubernetes with MLflow/Kubeflow; added canary rollouts and automated model versioning.

• Deployed evaluation on AWS SageMaker; cut model training costs ~25% via spot strategies and optimized instances.

• Implemented LLM-powered clinical note summarization and eligibility checks using prompt templates and RAG/LangChain.

• Established model risk controls, bias checks, SHAP/LIME explanations & performance drift monitors with on-call playbooks.

• Orchestrated PySpark feature pipelines to PostgreSQL; improved data freshness from daily to hourly.

• Partnered with RA/QA to document validation & traceability for FDA-aligned releases; closed audit findings in first pass. Data Scientist/ ML Engineer Coforge Feb 2021 - Jul 2023

• Delivered supervised models (XGBoost/Random Forests) for churn, risk, & propensity; improved AUC by 15% over baselines.

• Built NLP solutions for ticket triage and document classification; lowered manual routing effort by ~50%.

• Engineered scalable PySpark ETL on Airflow; created reusable feature stores and SLAs for data quality and lineage.

• Containerized services with Docker and deployed on Kubernetes; implemented HPA and blue/green releases.

• Set up MLflow for experiment tracking, model registry, and approvals; enforced CI/CD gates in Git-based workflows.

• Implemented model monitoring (latency, drift, data skew) with Prometheus/Grafana; reduced MTTR for incidents by ~30%.

• Stood up analytics dashboards in Power BI/Tableau to operationalize KPIs and explain model lift to business stakeholders.

• Built REST inference APIs with FastAPI/Flask; added auth, rate-limits, and request logging for secure operations.

• Piloted LLM/RAG POCs for contract Q&A and summary generation using FAISS/Pinecone and governance guardrails.

• Drove Agile delivery (sprint planning, demos, retros) and wrote clear design docs/SOPs for handoffs to support teams. PROJECTS

Multi-Agent Generative Model for Personal Finance & Investment Guidance

• Built multi-agent LLM system for budgeting, stock analysis, and sentiment extraction, coordinated by central manager agent.

• Integrated RAG pipelines with Hugging Face and deployed via Streamlit, enabling interactive investment recommendations. Research Paper Summarization Model

• Designed hybrid classification + generative model summarizing academic texts based on depth, tone, and audience needs.

• Implemented LLMs with embeddings and semantic filtering, allowing real-time customizable summaries for diverse users. Ai - restaurant desk attendant (Voice to voice model project)

• Developed LLM-based voice ordering system trained on 400+ menus, handling automated customer orders in real time.

• Converted speech to structured orders generating receipts, kitchen notifications, and bills, improving accuracy by 25%. EDUCATION

• Masters of Professional Studies in Data Science

University of Maryland Baltimore County (UMBC) Maryland, USA

• Bachelor of Engineering in Computer Science

R.V.R & J.C College of Engineering, India

CERTIFICATIONS

• AWS Cloud Solution Architect - Professional

• Data analytics by python

• Problem solving through programming in C

• Data analysis with R

• Google Analytics

Contact this candidate