⭐️ What is Voximplant?
Voximplant is a voice first CPaaS platform processing more than 170M calls per month for 30,000+ customers worldwide. Our solutions help developers and product leaders who want to develop custom communications experiences quickly and without sacrificing future flexibility. Voximplant Platform is the most powerful communications platform that enables you to add voice, video, messaging, natural language processing and Gen AI to your applications. The Voximplant Platform provides an integrated communications development experience that includes native serverless tools, many AI integrations, and a comprehensive set of APIs and SDKs.
⭐️ Key Responsibilities
Design and Optimization of Prompts: Create, test, and continuously refine complex prompt systems for voice bots, including:
Conversational Flows: Designing multi-turn dialogue scenarios that feel natural and intuitive for the user.
Role-Playing Scenarios: Crafting precise persona and role-based prompts to ensure the bot adheres to its designated character (e.g., customer support agent, sales representative).
Data Extraction: Developing prompts to accurately identify and extract key information from user speech.
Dynamic Prompt Management: Implement and manage logic for the dynamic adaptation of prompts in real-time based on conversation context, user sentiment, and specific variables.
Function Calling \& Variable Integration: Design and integrate robust systems using function calling and variables to fetch external data, execute actions, and personalize interactions within the conversation flow.
RAG Implementation: Utilize and optimize Retrieval-Augmented Generation (RAG) systems to provide the bot with access to specific, up-to-date knowledge bases. Specifically, calling functions and existing RAG tools to enrich agent responses with up to date live data, and extracting key data from the conversation.
Testing and Quality Assurance: Rigorously test and evaluate LLM responses against criteria such as accuracy, relevance, completeness, safety, and brand tone alignment.
Analysis and Iteration: Analyze problematic interactions, user feedback, and performance metrics to identify root causes and implement prompt corrections and improvements.
Prompt is generally expected to be written in English.
⭐️ Required Qualifications \& Skills
2+ years of hands-on experience working with Large Language Models (LLMs), preferably with commercial models like GPT-4, Claude, or similar.
Proven experience (1+ years) in prompt engineering, specifically for developing and testing robust, multi-turn conversational agents (chatbots or voice bots).
Strong skills in testing prompts for functionality, reliability, and resilience to varied user inputs and re-generations.
Solid understanding and practical experience with LLM function calling to create dynamic, action-oriented bots.
Proficiency in working with variables and context management to maintain coherent and personalized conversations.
Basic understanding of programming (e.g., Python, JavaScript), working with APIs, and integrating with development tools.
Strong analytical skills to debug suboptimal LLM outputs and systematically improve prompt design.
Understanding of and experience with RAG (Retrieval-Augmented Generation) architecture.
⭐️ What Will Make You a Strong Candidate (Preferred Qualifications)
Experience with voice-specific platforms (e.g., Google Gemini, Ultravox, Amazon Lex, 11labs) or speech-to-text/text-to-speech technologies.
Familiarity with the cultural nuances, communication styles, and linguistic diversity of the Latin American market (e.g., understanding differences between Mexican, Colombian, and Argentine Spanish).
Knowledge of Portuguese is a significant plus for the Brazilian market.
A portfolio or examples of previous conversational AI projects you have contributed to.
⭐️ We provide
Best in class cloud communication platform
Transparent career growth opportunities
Company benefits
Remote location in LATAM
Job Types: Full-time, Permanent
Pay: Up to $52,000.00 per month
Work Location: Remote