Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Hi there! I'm Eryk Rudaski, an AI and Python backend engineer with a Master’s in Artificial Intelligence. I design scalable AI systems, specializing in ML, LLMs, and multi-modal AI, and I enjoy turning research ideas into production-ready solutions. Over the past nine years I’ve built and led AI solutions at major tech and research organizations, delivering measurable improvements in model accuracy, efficiency, and user engagement. I thrive in cross-functional teams, mentoring engineers and delivering robust cloud-based architectures.…Hi there! I'm Eryk Rudaski, an AI and Python backend engineer with a Master’s in Artificial Intelligence. I design scalable AI systems, specializing in ML, LLMs, and multi-modal AI, and I enjoy turning research ideas into production-ready solutions. Over the past nine years I’ve built and led AI solutions at major tech and research organizations, delivering measurable improvements in model accuracy, efficiency, and user engagement. I thrive in cross-functional teams, mentoring engineers and delivering robust cloud-based architectures.

Eryk Rudaski

Full Stack Developer, AI Engineer, AI Developer, +3





Hi there! I'm Eryk Rudaski, an AI and Python backend engineer with a Master’s in Artificial Intelligence. I design scalable AI systems, specializing in ML, LLMs, and multi-modal AI, and I enjoy turning research ideas into production-ready solutions. Over the past nine years I’ve built and led AI solutions at major tech and research organizations, delivering measurable improvements in model accuracy, efficiency, and user engagement. I thrive in cross-functional teams, mentoring engineers and delivering robust cloud-based architectures.…Hi there! I'm Eryk Rudaski, an AI and Python backend engineer with a Master’s in Artificial Intelligence. I design scalable AI systems, specializing in ML, LLMs, and multi-modal AI, and I enjoy turning research ideas into production-ready solutions. Over the past nine years I’ve built and led AI solutions at major tech and research organizations, delivering measurable improvements in model accuracy, efficiency, and user engagement. I thrive in cross-functional teams, mentoring engineers and delivering robust cloud-based architectures.

Available to hire

Hi there! I’m Eryk Rudaski, an AI and Python backend engineer with a Master’s in Artificial Intelligence. I design scalable AI systems, specializing in ML, LLMs, and multi-modal AI, and I enjoy turning research ideas into production-ready solutions.

Over the past nine years I’ve built and led AI solutions at major tech and research organizations, delivering measurable improvements in model accuracy, efficiency, and user engagement. I thrive in cross-functional teams, mentoring engineers and delivering robust cloud-based architectures.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Language

English

Fluent

Work Experience

Team Lead at Sage Software Inc

June 1, 2025 - October 1, 2025

Led and mentored a cross-functional team through the end-to-end full-stack development of a multi-agent Cryptocurrency AI assistant for market analysis and price forecasting, leveraging LangChain and LangGraph to coordinate agent-based reasoning, tool invocation, and conversational workflows in Python + React environment. Designed and operationalized an evaluation and observability pipeline using LangSmith and Guardrails AI, implementing an RLHF loop with PPO to improve response correctness and stability; achieved >95% acceptance accuracy on validated benchmarks. Deployed and scaled the system on AWS using SageMaker for model training and inference, MLflow for experiment tracking and model versioning, and CI/CD pipelines across staging and production environments. Built an end-to-end LLM multi-agent system with MCP and Microservice architecture, and enhanced sentiment analysis using the Santiment API. Optimized prompts and orchestration to reduce end-to-end latency by ~50% while preser

Staff AI Full Stack Engineer | Alex Product Team Lead at DriveHealth

October 1, 2023 - May 1, 2025

Led design and deployment of a HIPAA-compliant, multi-agent Voice AI assistant for clinical applications using LangChain/LangGraph to orchestrate reasoning and conversational flows over LiveKit. Architected a Hybrid Retrieval-Augmented Generation (RAG) system with PgVector and HNSW indexing, integrating metadata filtering and a TinyBERT reranker to reduce non-grounded responses by 35% and boost retrieval precision by 40%. Pioneered Generative AI Guardrails and safety checker models; deployed 3+ multi-modal RAG apps with React/FastAPI and PostgreSQL. Led distributed fine-tuning (Llama, Gemma) with PyTorch/QLoRA on SageMaker, accelerating convergence by 45% and reducing domain hallucinations by 40%. Implemented RLHF with PPO to improve diagnostic accuracy by 25%, and integrated MCP server with LiveKit for low-latency voice grounding. Developed HITL-based evaluation pipelines and monitoring via Prometheus/Grafana.

Senior Full Stack Engineer at Intellias

May 1, 2022 - October 1, 2023

Developed the Content Creator Multi-Agent System using LangChain/LangGraph with Neo4j, PostgreSQL, and Pinecone for RAG pipelines; integrated OpenAI GPT-4o and Tavily Search tools, achieving improved accuracy by 30%. Built price-prediction models (LSTM) with Kronos, doubling baseline accuracy; developed trading-signal generation with LLaMA and moving-average indicators; designed Snowflake / Azure Databricks data stack for data processing and scalable ML workflows.

Full Stack Engineer at Microsoft

July 1, 2017 - March 1, 2020

Built multivariate time-series forecasting using hybrid LSTM with attention, aligning with early Transformer-like architectures; fused heterogeneous data sources (chemical yields, lab analytics, supply-chain throughput, external market prices). Developed end-to-end MLOps pipeline (Kafka/Batch, feature store, PyTorch/TensorFlow, REST API + batch scoring, live back-testing); reduced forecasting error ~25% vs ARIMA/LSTM baseline. Mentored junior data scientists and chemical engineers on attention-based models and domain-chemistry features integrated with ML workflows.

Staff AI Full Stack Engineer at DriveHealth

October 1, 2023 - May 1, 2025

Led the design and deployment of a HIPAA-compliant, multi-agent Voice AI assistant for clinical applications, orchestrating complex reasoning and conversational flows over LiveKit with Python using LangChain + LangGraph. Architected and scaled a Hybrid Retrieval-Augmented Generation (RAG) system with PgVector and HNSW indexing, incorporating metadata filtering and a TinyBERT reranker to reduce non-grounded responses by 35% and boost retrieval precision by 40%. Pioneered Generative AI Guardrails and custom safety checker models (TinyBERT) to ensure regulatory compliance and prevent toxic outputs. Engineered and deployed 3+ multi-modal RAG applications using React, FastAPI, and PostgreSQL, including an admin dashboard for the calling center and a migration of React Router v5 → v7 with minimal downtime. Designed distributed fine-tuning pipelines for open-source LLMs (Llama, Gemma) using PyTorch, QLoRA, and instruction tuning on SageMaker, accelerating convergence by 45% and reducing dom

Senior AI & Python Backend Engineer | Technical Team Lead at DriveHealth

October 1, 2023 - May 1, 2025

Designed and deployed a HIPAA-compliant, multi-agent Voice AI assistant for clinical applications, orchestrating complex reasoning and conversational flows over LangChain/LangGraph with LiveKit in Python. Scaled a Hybrid Retrieval-Augmented Generation (RAG) system using PgVector and HNSW indexing, integrating a TinyBERT reranker to reduce non-grounded responses by 35% and boost retrieval precision by 40%. Built high-throughput, asynchronous backend services using FastAPI, asyncio, and gRPC to handle real-time voice sessions, LLM inference, RAG retrieval, and safety checks, sustaining sub-300 ms end-to-end latency under concurrent clinical workloads. Implemented secure HIPAA-compliant APIs and data pipelines with PostgreSQL, Redis, and AWS (S3, IAM, KMS) with RBAC and audit logging. Led distributed fine-tuning pipelines for open-source LLMs (Llama, Gemma) using PyTorch, QLoRA and Instruction Tuning on SageMaker, accelerating convergence and reducing domain-specific hallucinations. Imple

Senior Python Backend Engineer at Intellias

May 1, 2022 - October 1, 2023

Developed Content Creator Multi-Agent System using LangChain and LangGraph with KnowledgeBase (Neo4j, PostgreSQL) and vector stores (Pinecone) for a robust RAG pipeline, incorporating OpenAI, GPT-4o, Tavily Search tools, and semantic search to improve accuracy by 30%. Built and optimized a price-prediction model using LSTM and Kronos, achieving a 2x improvement, and developed a trading signal generation algorithm using LLaMA with multiple indicators, improving accuracy by ~20%. Constructed data pipelines and analytics environments in Snowflake and Azure Databricks to support ML workflows and auto-scaling.

Backend Engineer Onsite at Microsoft

July 1, 2017 - March 1, 2020

Designed and scaled Python-based microservices for Azure Cognitive Services (Speech & Text Analytics), optimizing async request pipelines with aiohttp and gRPC to reduce end-to-end inference latency by 35%. Built distributed data ingestion and preprocessing for Office 365 telemetry via Azure Event Hubs, Azure Functions (Python), and Cosmos DB, processing hundreds of millions of events per day with strict schema evolution and backward compatibility guarantees. Engineered high-availability backend services for Bing personalization and ranking signals, integrating Python feature extraction pipelines with Azure Data Lake, Spark, and Redis caching layers, improving feature freshness SLAs from hours to sub-10-minute intervals. Implemented secure, multi-tenant backend APIs for Dynamics 365 with Python, ASP.NET interop layers, OAuth 2.0, and Azure Active Directory, enforcing RBAC and data isolation across thousands of enterprise tenants. Automated CI/CD and infrastructure provisioning for Pyth