MOHITH CHOWDARY Surabathuni

Fulshear,TX

Summary

AI professional with proven track record in developing cutting-edge artificial intelligence systems. Adept at designing and implementing machine learning models that solve complex problems. Focused on team collaboration and delivering reliable solutions that adapt to changing requirements. Expertise in neural networks and data analysis.

Professional with strong background in artificial intelligence and machine learning. Skilled in developing, implementing, and optimizing AI algorithms and models. Proven track record of enhancing team collaboration and delivering impactful results. Adaptable, reliable, and ready to meet evolving project demands.

Overview

years of professional experience

Work History

AI ENGINEER

LOGIC 20/20

Seattle, WA, USA

01.2023 - Current

Designed and deployed an AI-powered knowledge platform using Retrieval-Augmented Generation (RAG) integrated with Azure OpenAI, Amazon SageMaker, and Vertex AI to ground LLM responses in proprietary enterprise documents across multi-cloud environments (AWS, Azure, GCP).
Built end-to-end content preparation pipelines including document ingestion, chunking, semantic enrichment, OCR integration, and metadata indexing to improve retrieval quality, contextual relevance, and downstream LLM answer accuracy.
Leveraged Databricks (MLflow, Genie model) for distributed training, experiment tracking, model registry management, and automated deployment of regression, time-series forecasting, and deep learning models (CNNs, RNNs, LSTMs).
Designed and implemented LLM-powered RAG pipelines including tokenization strategies, model selection, fine-tuning patterns (LoRA, PEFT), and embedding workflows integrated with SQL and NoSQL databases.
Developed hybrid retrieval systems combining keyword search, vector similarity search, and semantic ranking to improve recall and precision for conversational and vague enterprise queries.
Implemented agentic workflows and multi-agent orchestration using Model Context Protocol (MCP), enabling tool calling, query decomposition, parallel subqueries, and secure integration with enterprise APIs and databases.
Built scalable AI inference endpoints using FastAPI and Docker, deploying on Azure Kubernetes Service (AKS), Google Kubernetes Engine (GKE), and AWS EC2 with container images managed through ACR, ECR, and Artifact Registry.
Implemented serverless execution patterns using AWS Lambda, Azure Functions, and Google Cloud Run to decouple ingestion, embedding generation, and re-indexing workflows from user-facing request paths.
Secured APIs and model endpoints using Azure Active Directory (Entra ID), AWS IAM, Google Cloud IAM, OAuth2/JWT authentication, API Gateway, Azure API Management, and centralized secrets management with RBAC and credential rotation.
Built CI/CD pipelines using GitHub Actions to automate unit testing, container builds, model packaging, versioning, and staged deployments across AWS, Azure, and GCP production environments.
Operationalized monitoring and observability using CloudWatch, Azure Monitor, and Google Cloud Monitoring with structured logging, distributed tracing, alerting, inference telemetry collection, and drift detection.
Optimized system latency and operational cost through embedding batching, caching strategies, top-k tuning, relevance thresholds, evaluation harness development (groundedness, relevance, citation coverage), and performance-aware development practices.
Environment: Multi-cloud Generative AI platform across AWS (SageMaker, Lambda, EC2, S3, ECR, API Gateway, IAM, CloudWatch), Azure (Azure ML, Azure OpenAI, Functions, AKS, Blob Storage, API Management, Entra ID, Azure Monitor), and GCP (Vertex AI, Cloud Run, GKE, Cloud Storage, Artifact Registry, IAM, Secret Manager, Cloud Monitoring). Databricks (AutoML, MLflow, model lifecycle automation, Genie), regression models, time-series forecasting, CNN/RNN/LSTM architectures, LoRA, PEFT, prompt engineering, LLM latency and performance evaluation, Python (pandas, numpy, scikit-learn, PyTorch, TensorFlow), Seaborn visualization, feature engineering, Docker, FastAPI, serverless functions, CI/CD (GitHub Actions), Streamlit, HTML/CSS/JavaScript, vector databases (FAISS, Pinecone), hybrid retrieval, SQL/NoSQL databases, monitoring, distributed tracing, and model drift detection.

ML Engineer

Capgemini

Hyderabad, India

12.2019 - 06.2022

Built an Integration Copilot that converted natural-language requests into validated API calls, integrating enterprise services across AWS and Azure to deliver secure, production-grade AI-enabled automation.
Developed and deployed ML services using FastAPI and Docker, running on Azure Virtual Machines, Amazon EC2, and Google Compute Engine with containerized workloads managed through ACR, ECR, and Artifact Registry.
Implemented serverless compute patterns using AWS Lambda, Azure Functions, and Google Cloud Functions for asynchronous inference and background data processing.
Engineered secure API layers using Amazon API Gateway, Azure API Management, OAuth2/JWT authentication, Azure Active Directory, AWS IAM, and Google Cloud IAM to protect enterprise ML endpoints.
Built end-to-end ML pipelines including advanced data cleaning, feature engineering, transformation, and visualization using pandas, NumPy, and Seaborn.
Integrated relational SQL databases and NoSQL datastores with document-style retrieval systems to support enterprise automation and analytics workflows.
Automated CI/CD workflows using GitHub Actions and Jenkins for model packaging, container publishing, and environment-based deployments.
Implemented monitoring and observability using AWS CloudWatch, Azure Monitor, and Stackdriver (Google Cloud Monitoring) to track latency, performance, and model drift.
Built orchestration workflows using messaging systems and background job frameworks to support reliable, scalable inference workloads across multi-cloud environments.
Applied systems architecture principles to ensure secure networking (private endpoints), infrastructure stability, and seamless integration of ML services into downstream enterprise systems.
Maintained production governance through RBAC policies, centralized secrets management, audit logging, and structured release workflows across AWS, Azure, and GCP platforms.
Environment: Multi-cloud ML platform across AWS (SageMaker, EC2, S3, Lambda, API Gateway, IAM, CloudWatch), Azure (Azure ML, Functions, Virtual Machines, Blob Storage, API Management, Active Directory, Azure Monitor), and GCP (Vertex AI AutoML, Compute Engine, Cloud Storage, Cloud Functions, Cloud IAM, Stackdriver). Databricks with MLflow for distributed training and experiment tracking, regression modeling, time-series forecasting, CNN/RNN/LSTM architectures, Python (pandas, NumPy, scikit-learn, PyTorch, TensorFlow), Seaborn visualization, FastAPI, Docker, serverless runtimes, CI/CD pipelines, SQL/NoSQL databases, OAuth2/JWT authentication, RBAC, secrets management, data engineering, model monitoring, drift detection, distributed tracing, and production ML lifecycle management.

Education

Master’s - computer and information systems

Rivier University

Nashua, New Hampshire, USA

12.2024

Skills

Generative AI & Deep Learning:
Proficient in transformer architectures
Techniques: Prompt Engineering, Fine-tuning, Transfer Learning, RAG (Retrieval-Augmented Generation), Few-shot and Zero-shot Learning, LoRA, PEFT (parameter-efficient fine-tuning)
Applications: Text Generation, Image Synthesis, Conversational AI, Code Generation, Synthetic Data across distributed cloud infrastructures
Deep Learning: PyTorch, TensorFlow, Keras, JAX with CNN/RNN/LSTM architectures trained and deployed in multi-cloud production environments
NLP: Hugging Face Transformers, spaCy, NLTK, LangChain for enterprise NLP systems across AWS, Azure, and GCP
CV & Multimodal: OpenCV, DALL-E, CLIP, Stable Diffusion integrated with cloud-native storage and compute services
Experiment Tracking: Weights & Biases, TensorBoard, MLflow, Databricks MLflow for experiment tracking, model lifecycle automation, and reproducible workflows
Vector Databases: FAISS, Pinecone, Weaviate, ChromaDB supporting hybrid retrieval and vector search across multi-cloud architectures

Data Processing: Pandas, NumPy, Polars, Dask, PySpark for distributed data engineering and feature engineering at scale
Data Visualization: Seaborn, Matplotlib for model evaluation and performance analytics
Workflow Orchestration: Airflow, Prefect for automated training, evaluation, and retraining pipelines
Model Deployment: ONNX, TorchServe, TensorFlow Serving, FastAPI with Docker, Kubernetes, and serverless functions
Cloud Platforms: AWS (SageMaker, EC2, Lambda), Azure (Azure ML, Functions, Virtual Machines), GCP (Vertex AI, Cloud Run) with multi-cloud architecture design
Platform & MLOps: Databricks (AutoML, MLflow, model lifecycle automation), CI/CD Pipelines, Model Monitoring, Model Version Control, LLM latency and performance evaluation
DevOps Tools: Docker, Kubernetes, GitHub Actions, Jenkins across multi-cloud CI/CD environments
Core Languages: Python, JavaScript (Nodejs), Bash
Others: SQL, C, TypeScript (frontend integration), multi-cloud architectures (AWS, Azure, GCP), regression modeling, time-series forecasting, CNN/RNN/LSTM architectures, Streamlit, serverless functions, anomaly detection, predictive modeling, performance analytics, AI governance, systems architecture, data storage, Snowflake, scripting, security engineering, Apache ecosystem, ETL pipelines, NoSQL databases, object storage, large language models, feature engineering, data engineering, version control systems (Git), VSCode, JupyterLab, data structures, computer vision, linear algebra, scikit-learn, algorithms, calculus, distributed databases, tensor libraries, numerical computing, API development, MCP integration, NLP systems

Websites

Timeline