Prompt Engineer

Prompt Engineer — Aetheria Intelligence

Boston, MA | January 2024 – Present

• Developed 24 system prompt templates for enterprise RAG agents using LangChain and Python, reducing latency by 28% and eliminating hallucinated answers across 15,000 monthly user sessions.
• Built automated evaluation pipelines using LangSmith and DSPy to benchmark LLM outputs, raising compliance with JSON output schemas from 74% to 98.6%.
• Optimized prompt templates for GPT-4 and Claude 3.5 Sonnet to perform few-shot categorization, decreasing API token consumption by 32% and saving $14,000 in monthly operational costs.
• Implemented prompt-injection guardrails and input-filtering policies using LlamaGuard, decreasing successful jailbreak attempts by 99.4% during penetration testing.

AI Interaction Specialist — Cognitive Stream Technologies

Austin, TX | July 2020 – December 2023

• Designed few-shot and chain-of-thought prompting workflows for customer support chatbots, which resolved 42% of incoming tickets without agent intervention.
• Structured prompt context boundaries and system instructions to handle long-document summarization, increasing data ingestion accuracy by 25% across 500+ client dossiers.
• Coordinated with a team of 6 backend developers to integrate LLM API endpoints into existing Django microservices, accelerating feature deployment cycles by 18%.
• Conducted A/B testing on 15 prompt variants for marketing copy generation, which boosted click-through rates by 12% and improved text relevance scores by 35%.

Prompt engineering, Large Language Models (LLMs), Chain-of-Thought prompting, Few-shot learning, Context engineering, Retrieval-Augmented Generation (RAG), AI agent workflows, DSPy, LangChain, LangSmith, Python, LlamaIndex, Vector databases, Pinecone, ChromaDB, Prompt vulnerability testing, Prompt security, LLM evaluation, JSON schema validation, Semantic search, API integration

Enterprise Semantic Agent Platform

Role: Lead Prompt Designer

Tools: Python, LangChain, ChromaDB, Claude 3

Engineered RAG-based context injection templates that indexed 10,000 internal documents, reducing customer service response time by 45% with 0% critical hallucinations.

Automated Prompt Evaluation Suite

Role: AI Specialist

Tools: LangSmith, DSPy, OpenAI API, GitHub Actions

Created a continuous evaluation pipeline for prompt upgrades, which cut production regression testing duration from 4 hours to 10 minutes and caught 18 formatting errors before deployment.

• Vanderbilt University Prompt Engineering for ChatGPT (2023)
• DeepLearning.AI ChatGPT Prompt Engineering for Developers (2023)
• AWS Certified AI Practitioner (2025)
• Microsoft Certified: Azure AI Fundamentals (2022)

• Languages: English (Native), Ukrainian (Fluent)
• Open Source: Contributor to DSPy and LangChain prompt templates repository
• Availability: 2 weeks notice