Skills & Credentials

LLM & Agent Development

Frameworks & Providers: LangChain, LangGraph, CrewAI, OpenAI, Anthropic, Claude

Capabilities: Retrieval-Augmented Generation (RAG), Agent / Agentic Workflows, Multi-Agent Orchestration, Model Routing, Tool Calling / Function Calling, Structured Output Generation, Context Engineering, Human-in-the-Loop Checkpointing

Architecture Patterns: Deterministic supervisors over non-deterministic workers, agent handoff and shared state, failure isolation, inference cost as a design constraint

Model Serving & Delivery

Inference & Endpoints: vLLM, NVIDIA Triton Inference Server, Ollama, self-hosted and air-gapped serving, hybrid routing between frontier and local models, streaming agent endpoints on FastAPI and FastMCP, SageMaker endpoints

Evaluation & Observability: Langfuse (self-hosted), tracing, token cost telemetry, run replay

Platforms: Amazon Bedrock, SageMaker Studio, MLflow, Databricks, Amazon Nova, Amazon Q, Amazon Titan embeddings, sentence-transformers

Front-End: React, Vite, SCSS, Bootstrap

IaC & Automation: Terraform, AWS CDK, CloudFormation, Ansible, GitHub Enterprise

Orchestration: Kubernetes, GitOps reconciliation, container image bakes

Retrieval, Search & Vectors

OpenSearch: Neural Search, BM25, Semantic Search, Hybrid Search, Search Pipelines, RAG Processor, ML Inference, Ingest Pipelines

Retrieval Design: Chunking and embedding strategy, reranking, semantic deduplication, vector store selection tradeoffs

Vector Stores: S3 Vectors, OpenSearch k-NN, embedding model selection and dimension tradeoffs

AI / ML

Computer Vision: Image and video labeling, object and scene detection, face and in-image text detection

Document AI & OCR: OCR extraction, form and table parsing, document classification and routing

Speech & Language: Entity extraction, sentiment and topic modeling, translation

Prediction: Anomaly detection, forecasting, recommendation, fraud scoring

MLOps & Pipelines: SageMaker Pipelines, MLflow model registry, training and tuning jobs, batch and real-time inference endpoints

Practice: Feature engineering, dataset labeling, train / eval splits, drift monitoring, model evaluation metrics

Data Engineering

Processing: Apache Spark, Kinesis Data Streams, Glue, Athena, Amazon Managed Workflows for Apache Airflow (MWAA), SQS-decoupled event pipelines

Storage & Lakes: Iceberg Data Lake, S3 + Parquet (Hive-partitioned), S3 Vectors

Pipeline Design: Zero-loss migration cutovers, schema evolution, partitioning strategy, replay and backfill

Cybersecurity & Analytics

Security Tools: Splunk, SailPoint, CrowdStrike, Graylog, Tenable, Proofpoint, Shodan

Frameworks: MITRE ATT&CK, NIST, CIS Controls, GDPR

Practices: Security Information and Event Management (SIEM), Threat Intelligence, Threat Hunting, Cyber Investigations, Open-Source Intelligence (OSINT), Fraud Detection, Attack-Surface Enumeration, OSINT Collection Pipelines

AWS Cloud Platform

AI/ML: SageMaker, Bedrock, Fraud Detector, Comprehend, Personalize, Textract, Rekognition, Transcribe, Translate, Amazon Q

Data & Analytics: Athena, Glue, Kinesis, Amazon Managed Streaming for Apache Kafka (MSK), OpenSearch Service, Data Lakes, S3 Vectors, Amazon Managed Workflows for Apache Airflow (MWAA)

DevOps & Integration: CDK, CloudFormation, AWS CI/CD, Step Functions, ECR, Lambda, IAM, EKS, API Gateway, EventBridge, SQS

Platform: Systems Manager, ELB, KMS, Organizations, Control Tower, Amazon Connect

Security: Security Hub, GuardDuty, Inspector, Macie