Machine Learning Engineer

Sumith Reddy Likkidi

Machine Learning Engineer

PROFILE SUMMARY:

• 6+ years of experience in designing, developing, and deploying machine learning models, artificial intelligence solutions, and advanced analytics for real-world applications and expertise in Python, R, Java, and Scala to develop scalable machine learning pipelines and data workflows for high-performance applications.

• Hands-on experience with frameworks and libraries like TensorFlow, PyTorch, Scikit-learn, Kera’s, OpenCV, and XGBoost for model building and optimization and Strong capabilities in data cleaning, preprocessing, feature engineering, and dimensionality reduction using Pandas, NumPy, and Spark to enhance data quality for analysis.

• Well-versed in Hadoop, Apache Spark, and Hive for distributed data processing and large-scale model training in big data environments and extensive experience working with cloud platforms including AWS (Sage Maker, Lambda), Google Cloud (Vertex AI, Big Query), and Azure (Machine Learning Studio) for cloud-based machine learning deployments and skilled in deploying machine learning models into production environments using Docker, Kubernetes, and REST APIs for seamless integration and expertise in MLOps practices and tools like Git, Jenkins, and CI/CD pipelines to automate workflows and ensure reproducibility of machine learning models.

• Strong experience in data visualization with tools such as Tableau, Power BI, and Matplotlib/Seaborn to present actionable insights to stakeholders and Proficient in building real-time streaming applications using Ka:a, Flink, and Spark Streaming to process high-throughput data.

• Advanced knowledge in graph-based machine learning and network analysis using NetworkX and PyTorch Geometric for complex data structures and experienced in building and fine-tuning language models such as BERT, GPT, and XLNet for a wide range of NLP applications and expertise in edge AI deployment for resource-constrained environments with tools like TensorFlow Lite and ONNX Runtime for efficient model deployment on mobile and IoT devices.

• Proficient in integrating machine learning models into production systems using frameworks like Flask and Fast API to develop scalable APIs for end-users and Proficient in time-series forecasting using ARIMA, SARIMA, and deep learning models such as LSTMs and GRUs to predict future trends in financial, healthcare, and sales data.

• Skilled in using natural language generation (NLG) libraries such as OpenAI GPT-3, T5, and BART for automating content generation and summarization tasks and adept at deploying models in production environments with Kubeflow, MLflow, and TensorFlow Serving for seamless model serving and management.

• Experience with knowledge graph construction and ontology modelling using Neo4j, Graph DB, and SPARQL to extract actionable insights from relational and non-relational data sources and expertise in building and deploying chatbots and virtual assistants using Dialog flow, Rasa, and Bot press for enhanced customer interaction and support.

• Skilled in working with big data tools like Apache Hadoop, Apache Flume, and Apache Ka:a to stream, process, and analyse data in real-time and Familiar with AutoML platforms like Google AutoML, H2O.ai, and TPOT to automate model selection, hyperparameter tuning, and feature engineering processes and Proficient in additional technologies such as SQL, NoSQL databases (MongoDB, Cassandra), and containerization tools like Docker and Kubernetes for efficient model deployment, as well as experience with web frameworks (Flask, Fast API) for building scalable APIs, and version control with Git to manage collaborative development and ensure code quality in team environments.

• Built automated ML pipelines integrating data ingestion, model training, and evaluation using Kubeflow, MLflow, and Airflow, enabling reproducible, end-to-end workflows in production environments.

• Built explainable AI (XAI) frameworks using tools like SHAP, LIME, and Captum to provide transparency and interpretability of model decisions for stakeholders in regulated industries. EDUCATION: University of New Mexico, Albuquerque, NM. 858-***-**** *************@*****.***

Experienced Machine Learning Engineer with 6+ years of expertise in designing, deploying, and optimizing data-driven solutions across manufacturing, finance, and insurance domains. Proven ability to build scalable machine learning pipelines, automate predictive analytics workflows, and translate business challenges

TECHNICAL SKILLS:

• Programming Languages: Python, R, Java, C++, Scala, SQL

• Machine Learning Frameworks: TensorFlow, Kera’s, PyTorch, Scikit-learn, XGBoost, LightGBM

• Deep Learning: CNN, RNN, GANs, BERT, GPT-3.

• Natural Language Processing (NLP): BERT, GPT, XLNet, spacey, NLTK, Genism, Text Blob, Word2Vec, Fast Text

• Data Preprocessing: Pandas, NumPy, Matplotlib, Seaborn, SciPy, OpenCV

• Big Data & Distributed Computing: Apache Hadoop, Apache Spark, Apache Kafka, Google Dataflow, Databricks

• Model Deployment & Integration: Docker, Kubernetes, Flask, Fast API, TensorFlow Serving, ONNX Runtime

• Cloud Platforms: AWS (Sage Maker, Lambda, EC2, S3), Google Cloud AI (Vertex AI, Big Query), Azure Machine Learning Studio

• Version Control & CI/CD: Git, GitHub, GitLab, Jenkins, Docker Compose, Kubernetes (CI/CD pipelines)

• ML Operations (MLOps): Kubeflow, MLflow, TensorFlow Extended (TFX), Data Robot

• Data Pipelines & ETL: Apache Airflow, Apache NiFi, AWS Glue, Talend, Informatica, Google Cloud Dataflow

• Computer Vision: OpenCV, YOLO, ResNet, Mask R-CNN, Object Detection, Semantic Segmentation

• Data Storage and Management: MySQL, PostgreSQL, MongoDB, Cassandra, HBase, NoSQL

• Parallel & Distributed Computing: Apache Spark, Dask, Horovod, TensorFlow Distributed WORK EXPERIENCE

Los Alamos National Laboratory Los Alamos, New Mexico Machine Learning Engineer Aug 2024 – Present

Description: Los Alamos National Laboratory (LANL) is a premier U.S. federal research facility historically critical site for nuclear development but a modern powerhouse in national security science, research, advanced facilities. MLEs take data science prototypes and develop them into functional machine learning models, ready for use in real-world applications and prepare and manage data, including data preprocessing, feature engineering, and data quality checks. Responsibilities:

• Optimized machine learning algorithms for scientific simulations and nuclear risk modelling, enhancing predictive accuracy in national security and energy research initiatives.

• Processed and analysed petabyte-scale scientific datasets using distributed computing frameworks like Hadoop and Apache Spark, accelerating complex physics computations and model training.

• Utilized Git, GitHub, and Bitbucket for collaborative research workflows, ensuring secure and traceable development of AI-driven modelling tools for experimental physics.

• Automated data ingestion pipelines from diverse scientific sources using Apache NiFi and Python, enabling real-time monitoring of experimental data from particle accelerators and fusion experiments.

• Developed interactive visualizations using Tableau and Power BI to communicate results from simulations and sensor arrays to interdisciplinary teams, enhancing decision-making in research programs.

• Built real-time machine learning pipelines using Apache Kafka and streaming analytics to monitor critical infrastructure and detect anomalies in facility operations and sensor data.

• Developed and optimized models using TensorFlow, PyTorch, and Scikit-learn for applications in materials science, plasma physics, and biosecurity threat detection, leveraging AWS and GCP for scalable training.

• Implemented CI/CD pipelines with Jenkins to automate the deployment of scientific models into high-security research environments, improving reproducibility and audit compliance.

• Managed configuration and infrastructure with Ansible to maintain consistency across high-performance computing clusters used in weapons simulations and climate modelling.

• Designed and maintained data schemas in MySQL, MongoDB, and PostgreSQL for organizing experimental and simulation datasets, enabling efficient querying and cross-study analytics.

• Containerized applications using Docker for consistent deployment of research tools across development and HPC environments, and orchestrated workflows with Kubernetes for scalable compute management.

• Leveraged OpenMP, MPI, and CUDA to parallelize scientific computing workloads on HPC clusters, reducing simulation runtime for nuclear material behaviour by over 60%.

• Integrated simulation outputs and sensor data using Apache Airflow and Talend into a centralized data pipeline, supporting multi-physics experiments and predictive modelling and Deployed LLMs such as BERT and GPT-3 on classified research text for secure document summarization, entity recognition, and automated knowledge extraction.

• Implemented Zero Trust Architecture and secure API gateways with OAuth2.0, Keycloak, and NGINX to protect sensitive AI pipelines in compliance with DOE security protocols. Environment: Hadoop, Apache Spark, Git, GitHub, Bitbucket, Apache NiFi, Python, Tableau, Power BI, Apache Kafka, TensorFlow, PyTorch, Scikit-learn, CI/CD pipelines, Jenkins, Docker, MySQL, MongoDB, PostgreSQL, Kubernetes, Apache Airflow, Talend, LLMs, BERT, GPT-3, OAuth2.0, Keycloak, NGINX. Kelly Cable (Cognizant) Albuquerque, New Mexico

Machine Learning Engineer Jan 2023 – Jul 2024

Description: Kelly Cable of New Mexico is a leading telecommunications and utility infrastructure company Specializing in the design, installation, and maintenance of underground and aerial fibre optic networks. Choosing and implementing appropriate machine learning algorithms based on the problem and data and developing and training machine learning models, including fine-tuning hyperparameters and evaluating performance. Responsibilities:

• Designed and deployed predictive machine learning models to optimize infrastructure planning, cable fault detection, and maintenance scheduling using TensorFlow, Scikit-learn, and PyTorch.

• Automated detection of anomalies in fibre optic network signals using real-time data processing frameworks like Apache Kafka, Apache Flink, and Spark Streaming, enabling rapid identification of service disruptions.

• Developed and monitored ETL pipelines for network telemetry data and performance logs using Apache Airflow, SQL, and Python, enhancing reporting accuracy and operational efficiency.

• Processed large volumes of geospatial and construction project data using Hadoop, Apache Spark, and HBase to support fibre deployment strategies and route optimization and built and deployed containerized applications using Docker and managed orchestration with Kubernetes for real-time fibre diagnostics and field technician tracking.

• Implemented network analytics dashboards using Tableau, Power BI, and Plotly to visualize uptime metrics, service coverage, and infrastructure performance KPIs. developed NLP models using BERT and GPT to classify and prioritize customer support tickets and automate telecom issue resolution for field teams.

• Leveraged XGBoost and LightGBM to forecast service demand and predict potential network congestion, enabling proactive capacity planning and equipment allocation and managed telecom asset and construction data using MongoDB, PostgreSQL, and Pandas, integrating information from project management and GIS systems.

• Applied reinforcement learning to optimize routing decisions for underground fibre installations, reducing cost and improving deployment timelines and ensured compliance with data privacy and infrastructure security policies across deployments, aligning with federal and industry-specific telecom standards.

• Engineered IoT-based monitoring systems using MQTT, EdgeX Foundry, and Raspberry Pi to track real-time signal strength and environmental factors affecting fibre network reliability and built automated network health classification models using AutoML, H2O.ai, and TPOT to assess line quality and proactively recommend maintenance schedules.

• Conducted time-series analysis using ARIMA, Prophet, and LSTMs to detect anomalies in network latency and predict cable degradation trends and developed custom RESTful APIs using Fast API and Flask to expose fibre infrastructure metadata and diagnostics for consumption by internal tools and external vendors. Environment: TensorFlow, PyTorch, Scikit-learn, NLP, BERT, GPT, Hadoop, Apache Spark, HBase, Pandas, NumPy, SQL, GDPR, CCPA, Tableau, Power BI, Plotly, Docker, Kubernetes, AI, GraphQL, REST APIs, XGBoost, LightGBM, MongoDB, PostgreSQL, Pandas, ARIMA, Prophet, LSTMs, Fast API, Flask. Presbyterian Healthcare Services Albuquerque, New Mexico Data Scientist Apr 2022 – Dec 2022

Description: Presbyterian Healthcare Services is a leading not-for-profit integrated health system offering a comprehensive range of medical services. Developing and implementing machine learning algorithms to build predictive models for tasks like disease diagnosis, treatment optimization, and patient outcome prediction Creating dashboards, reports, and other visualizations to communicate complex.

Responsibilities:

• Extracted, processed, and analysed large-scale patient records and clinical datasets using SQL, Python, and R to uncover health trends, improve patient care strategies, and support population health management.

• Developed and deployed predictive models with TensorFlow, PyTorch, and scikit-learn to forecast hospital readmission, predict chronic disease risk, and optimize care delivery workflows.

• Utilized Hadoop, Apache Spark, and Hive for processing electronic health records and real-time health monitoring data, enhancing clinical decision-making and operational efficiency.

• Designed interactive dashboards and clinical performance visualizations using Tableau, Power BI, and Matplotlib to deliver insights into care quality, patient outcomes, and resource utilization.

• Improved healthcare data architecture and storage solutions using Snowflake, Amazon Redshift, and Google Big Query, enabling secure and scalable access to patient and provider data.

• Performed spatial and demographic analysis using PostGIS, ArcGIS, and QGIS to evaluate healthcare access disparities and plan resource allocation across rural and underserved areas.

• Applied advanced Natural Language Processing (NLP) techniques with spacey, NLTK, and Hugging Face Transformers to extract insights from clinical notes, patient feedback, and health literature.

• Ensured development efficiency and deployment reliability using Git, GitHub, and Bitbucket with automated CI/CD pipelines through Jenkins, GitLab CI/CD, and CircleCI for healthcare data applications.

• Conducted sentiment analysis and clinical text mining using Text Blob, Genism, and BERT to monitor patient satisfaction, analyse provider feedback, and improve care experience. Environment: SQL, Python, R, TensorFlow, PyTorch, scikit-learn, Hadoop, Apache Spark, Hive, Tableau, Power BI, Matplotlib, Snowflake, Amazon Redshift, Google Big Query, PostGIS, ArcGIS, QGIS, spacey, NLTK, Git, GitHub, Bitbucket, Jenkins, GitLab CI/CD, CircleCI, Text Blob, Genism, BERT Shriram General Insurance Mumbai, India

Data Scientist / Machine Learning Engineer June 2019 – Dec 2021 Description: Shriram General Insurance Company Limited (SGIC) is a leading general insurance Specializing in motor, travel, personal accident, home, and commercial insurance. Building and training predictive models using statistical and machine learning techniques, such as regression, classification, and clustering and Deploying machine learning models into production environments and monitoring their performance. Responsibilities:

• Designed and implemented machine learning models for fraud detection, claim prediction, and underwriting risk assessment using TensorFlow, Kera’s, and PyTorch, enhancing operational accuracy and customer trust.

• Developed AI-driven solutions for policy recommendation systems, customer segmentation, and lapse prediction by utilizing advanced techniques in supervised, unsupervised, and reinforcement learning.

• Processed large volumes of policy, claim, and customer data using Apache Spark, Hadoop, and Kafka, enabling real-time analytics and risk profiling across the insurance lifecycle.

• Automated data ingestion, transformation, and cleansing using Apache Airflow and Apache NiFi, ensuring consistent and reliable data pipelines for actuarial and operational reporting.

• Applied Natural Language Processing (NLP) for extracting and summarizing data from claim forms, customer emails, and policy documents using spacey, BERT, and Hugging Face Transformers.

• Managed and queried both structured and unstructured insurance data using SQL, MySQL, PostgreSQL, and MongoDB for efficient analytics, storage, and reporting of customer and claim histories.

• Developed predictive models to reduce fraudulent claim payouts and identify high-risk policy applications, optimizing underwriting workflows with scikit-learn, XGBoost, and LightGBM.

• Containerized machine learning services using Docker and orchestrated deployments on Kubernetes, ensuring scalable and consistent environments for claims automation and risk modelling applications.

• Implemented CI/CD pipelines with Jenkins and GitLab CI/CD for automated deployment and testing of analytics models in production insurance systems.

Environment: SQL, Python, R, TensorFlow, PyTorch, scikit-learn, Hadoop, Spark, Hive, Tableau, Power BI, Matplotlib, Snowflake, Redshift, Big Query, spacey, NLTK, PostGIS, ArcGIS, MongoDB, Cassandra, Docker, Kubernetes, Jenkins.

Contact this candidate