Senior AI/ML Engineer for Shipable AI Systems

Parth Amruthbhai Patel Senior AI/ML Engineer

*******@*******.*** Cupertino, CA

SUMMARY

Senior AI/ML Engineer with 7+ years of experience designing, developing, and deploying scalable generative AI, agentic AI, and machine learning solutions across cloud-native environments. I build end-to-end AI platforms, LLM- powered agents, RAG pipelines, and intelligent APIs that actually ship : not just notebooks that look good. I collaborate closely with DevOps, backend, and product teams to turn business requirements into reliable, maintainable AI systems. I'm results-focused: I'd rather ship a working agent that automates one annoying task perfectly than a complex system nobody uses. Teamwork to me means unblocking others, documenting thoroughly, and being honest about trade-offs.

SKILLS

Generative AI & LLMs

GPT-4, GPT-4o, Claude 3.5 Sonnet, Gemini, Llama 2/3, Mistral, Hugging Face Transformers, LangChain, LangGraph, LlamaIndex, DSPy, RAG pipelines, embeddings (OpenAI, Cohere, Sentence Transformers), fine-tuning (LoRA, QLoRA, PEFT), prompt engineering, guardrails, structured outputs, few-shot prompting, prompt chaining Agentic AI

LangGraph, AutoGen, CrewAI, AWS Bedrock Agents, tool calling, function calling (OpenAI, Anthropic, Gemini), multi-agent workflows, supervisor/worker patterns, handoffs, hierarchical agents, human-in-the-loop (HITL), memory management (conversation buffers, semantic memory, vector stores) Vector Databases

Pinecone, Qdrant, Milvus, FAISS, ChromaDB, pgvector, Azure AI Search AI/ML Frameworks

PyTorch, TensorFlow, Keras, Scikit-learn, Hugging Face Transformers, ONNX, OpenCV, XGBoost, LightGBM Data Engineering & Processing

Apache Spark, Pandas, NumPy, Airflow, Kafka, ETL Pipelines, Feature Engineering, Data Validation MLOps & LLMOps

Docker, Kubernetes, Terraform, Jenkins, GitHub Actions, CI/CD Pipelines, MLflow, Kubeflow, ArgoCD, Helm, LangSmith, LangFuse, Arize Phoenix, Promptfoo, Git, Linux, Nginx Cloud Platforms

AWS (EC2, ECS, EKS, Lambda, S3, SageMaker, Bedrock, CloudWatch, IAM, API Gateway, RDS, Step Functions, SQS, DynamoDB), Microsoft Azure (AKS, OpenAI Service, AI Search), Google Cloud Platform (GCP, Vertex AI) Databases

PostgreSQL, MySQL, MongoDB, Redis, Elasticsearch

Backend & APIs

FastAPI, Flask, REST APIs, WebSockets, Microservices Architecture, Distributed Systems, API Integration Monitoring & Observability

Prometheus, Grafana, ELK Stack, Logging, Performance Monitoring Version Management

Git (trunk-based, bisect, reflog, cherry-pick, rebase), GitHub Actions, GitLab CI, semantic versioning, Git LFS, conventional commits

EXPERIENCE

Senior AI/ML Engineer

Cody Solutions - Austin, TX

05/2022 – 04/2026

•Designed and deployed enterprise-grade AI/ML platforms supporting generative AI, agentic AI, intelligent automation, document processing, predictive analytics, and large-scale NLP workloads.

•Built RAG pipelines for enterprise knowledge management using LangChain, Pinecone, and GPT-4. Implemented semantic search, vector embeddings, prompt orchestration, and citation tracking.

•Developed multi-agent systems using LangGraph and AutoGen with supervisor/worker patterns. Built specialized agents for database queries, document retrieval, API calls, and human approval workflows.

•Designed tool-calling schemas for internal APIs (Salesforce, Jira, Slack, AWS) with clear input schemas and safety constraints. Read-only operations ran autonomously; writes required human approval.

•Implemented human-in-the-loop (HITL) interfaces where users could see agent decision trees in real time, approve sensitive actions, and step in when agents were stuck.

•Built prompt versioning systems using Git and Promptfoo. Every prompt template lived in the repo with automated regression tests running on PRs.

•Set up LangSmith and LangFuse for agent tracing, debugging, and evaluation. The team could replay any agent conversation step by step to understand bad decisions.

•Implemented long-term memory for agents using PostgreSQL and vector embeddings. Agents remembered past interactions and could suggest known solutions to recurring problems.

•Built RAG evaluation pipelines using RAGAS and DeepEval. Measured faithfulness, answer relevance, and context recall on test sets of hundreds of questions.

•Developed production-ready RESTful APIs and AI microservices using FastAPI and Flask for internal and customer-facing applications.

•Implemented containerized AI workloads using Docker and Kubernetes (EKS) to improve deployment consistency and infrastructure scalability.

•Automated CI/CD workflows with GitHub Actions, Jenkins, and Terraform to streamline model deployment and infrastructure provisioning.

•Deployed machine learning and LLM workloads across AWS cloud infrastructure using EKS, Bedrock, SageMaker, Lambda, and S3 services.

•Collaborated with DevOps and backend engineering teams to improve monitoring, logging, and production reliability using Prometheus, Grafana, and ELK Stack.

•Developed scalable data processing systems using Apache Spark and Kafka for high-volume streaming and batch processing workloads.

•Improved model observability, versioning, and experiment tracking with MLflow and centralized logging systems.

•Participated in architecture reviews, technical planning, sprint execution, and mentoring junior engineers on Gen-AI and agentic AI best practices.

Software Developer / Machine Learning Engineer

Initialize - San Ramon, CA

07/2019 – 05/2022

•Started as a software developer and grew into machine learning and Gen-AI engineering over three years.

•Developed backend services and machine learning modules for data-driven business applications and workflow automation systems.

•Built LLM-powered automation tools – a tool that could read emails and draft responses, then moved to multi- step reasoning systems.

•Built a customer support triage agent: incoming tickets went into a queue, an LLM with tool access (database lookup, knowledge base search, ticket history) decided priority and suggested a response, with human approval before sending.

•Introduced LangGraph for agent workflows – classify ticket, search knowledge base, draft response. Checkpoints at each step let humans intervene when needed.

•Built and maintained REST APIs and backend services using Python and Flask within microservice-based environments.

•Assisted in training and deploying machine learning models for predictive analytics and customer behavior analysis.

•Created ETL pipelines and data preprocessing workflows using Pandas, NumPy, and SQL-based systems.

•Worked with cloud-hosted infrastructure and containerized deployments using Docker and AWS services.

•Collaborated with cross-functional teams to improve software performance, application stability, and deployment automation.

•Implemented Git-based version control workflows and participated in Agile software development processes.

•Developed internal automation tools and monitoring scripts to support system maintenance and operational efficiency.

•Supported integration of AI-driven features into existing enterprise applications and backend systems. CORE PROJECTS

•Enterprise AI Knowledge Assistant – Developed a large-scale AI assistant platform powered by LLMs and retrieval-augmented generation pipelines for enterprise knowledge management. Implemented semantic search, vector embeddings, prompt orchestration, and secure API integration using LangChain, FastAPI, PostgreSQL, and AWS infrastructure.

•Multi-Agent IT Operations Platform – Built a multi-agent system for IT automation with specialized agents for log analysis, database queries, ticket creation, and service restart (with human approval). Used LangGraph for orchestration, AWS Bedrock Agents for tool execution, and deployed on EKS with Terraform.

•Intelligent Document Processing System – Built an AI-driven document analysis platform capable of extracting, classifying, and validating structured data from business documents using OCR, NLP models, and transformer- based architectures. Deployed scalable inference services with Kubernetes and Docker.

•Predictive Analytics Pipeline – Designed an end-to-end machine learning pipeline for predictive business analytics, including data ingestion, feature engineering, model training, experiment tracking, and automated deployment workflows using MLflow, Airflow, and AWS SageMaker.

•LLM Playground for Internal Teams – Built a sandbox where engineers could test different models (GPT-4, Claude, Llama 3), adjust parameters, compare outputs side-by-side, and save presets to Git. Backend used vLLM for efficient inference. Used by dozens of engineers across the company.

•Real-Time Data Processing Platform – Developed distributed streaming pipelines using Apache Kafka and Spark for processing high-volume operational data. Integrated monitoring and observability tools to improve reliability and system visibility across production environments. EDUCATION

Bachelor's Degree in Computer Science

University of California Santa Cruz

08/2014 – 05/2018 Santa Cruz, CA

Contact this candidate