Backend Engineer
Full-timeAlternative military service eligible
Hiring process
- Document screening
- Résumé / portfolio in any format, as a PDF
- Email it to recruit@mindlogic.ai
- Interviews
- 1. Online interview → 2. In-person Tech & Fit interview
What you’ll do
Right after you join
- Design and build the backend APIs for our AI chat services. You’ll own core product domains: conversation handling, file and document processing (RAG), multimodal generation and credit billing.
- Abstract many LLM vendors behind a single gateway, connect each new model reliably as it ships, and measure its cost and quality.
- Work on multi-tenancy, SSO integration and permission models for university and enterprise customers — designing with on-premises and public-cloud environments in mind.
- Read traffic and cost metrics yourself, find bottlenecks and fix them. Run the whole cycle on your own: incident response, postmortems and the automation that keeps an issue from coming back.
- Raise each other’s bar through code and design reviews, and help shape the culture of the backend team and the wider engineering team.
Tech stack & tools
- Languages and frameworks: Python / FastAPI, TypeScript / Express
- Databases: PostgreSQL, Redis
- ORM: Tortoise ORM + Aerich, TypeORM
- Infrastructure: AWS (Elastic Beanstalk, EKS, Lambda, RDS, S3), NCP
- Communication: Jira, Slack, Notion
- AI tools: Claude Code, Cursor, Codex and more
Requirements
- 3+ years of backend development experience, or equivalent
- Solid server-engineering fundamentals (networking, databases, concurrency)
- Hands-on RDBMS schema design and query tuning
- Have owned a Python or TypeScript API server from design through operation
- Use AI coding agents (Claude Code, Codex, etc.) in your day-to-day development
- Have taken the initiative in an organization that builds and ships fast
- Have handled deployment, monitoring and incident response in the cloud
- Rather than just delivering the work you’re given, you find and analyze problems yourself and propose solutions to the team
Nice to have
- Have run LLM APIs in a production service (prompt design, token-cost optimization, streaming)
- Have built LLM agents or MCP tool servers yourself
- Experience with high-volume traffic or real-time processing (WebSocket, voice streaming)
- Have set up Kubernetes, IaC or CI/CD pipelines yourself
A message to future teammates
Hello from the Mindlogic backend team.
Our backend is not a simple CRUD server. We put dozens of LLMs behind a single gateway and serve them to many customers, build real-time voice and chat pipelines that keep long-term memory and personas, and bill usage down to the individual token. Models change and vendors overhaul their APIs on a very short cycle, so from the design stage we ask together, “How will this change six months from now?”
When something breaks, we find the root cause and automate it away so the same problem doesn’t come back. Rather than working through tickets, we aim to spot problems in metrics and logs first and propose fixes ourselves. AI coding tools are already standard on the team, so we spend less time on repetitive work and put the time saved into refining the architecture.
The product has grown fast, so there’s still plenty to improve. We’d love to work with someone who sees that as an opportunity.
Before you apply
- Any false information in your application, or found during the process, may end your candidacy
- Candidates eligible for employment protection are given preference under applicable law
AI Software Engineer
Full-timeAlternative military service eligible
Hiring process
- Document screening
- Résumé / portfolio in any format, as a PDF
- Email it to recruit@mindlogic.ai
- Interviews
- 1. Online interview → 2. In-person Tech & Fit interview
What you’ll do
- Build persona chatbots and applied services on LLM- and sLM-based AI models
- Build pipelines to optimize AI/ML model performance and bring models into production
- Design and implement RAG (Retrieval-Augmented Generation) and agent architectures
- Optimize the inference speed and quality of large language models
- Build AI response-generation systems for real-time streaming environments
- Follow the latest AI research and explore techniques we can bring into our services
- Advance persona grounding and long-term memory systems for personalized AI experiences
Requirements
- Bachelor’s degree or higher in computer science, AI or a related field, or equivalent hands-on experience
- 3+ years developing AI/ML projects and deploying them in commercial services
- Deep understanding of, and hands-on experience with, deep learning frameworks such as PyTorch and TensorFlow
- Understanding of LLMs, RAG and agent systems, with experience implementing them
- Deep understanding of prompt engineering and experience applying it in practice
- Proficiency in Python and related data processing and analysis libraries
- Ability to design and implement complex AI systems
- English good enough to read and write technical documents
Nice to have
- Experience building and optimizing LLM-based services
- Experience serving language models with vLLM, llama-cpp, SGLang or similar
- Experience with frameworks such as LlamaIndex and LangGraph
- Experience across AI modalities such as NLP, CV and speech recognition/synthesis
- Papers published in, or reviewing for, top venues (NeurIPS, ICML, ICLR, ACL, etc.)
- Awards in AI competitions in Korea or abroad (Kaggle, etc.)
- Experience building real-time conversational systems or streaming services
- Experience handling high traffic and optimizing performance
- Communicate well with other functions (product planning, design, etc.)
Tech stack
- Languages: Python, TypeScript/JavaScript
- Deep learning frameworks: PyTorch, TensorFlow, Hugging Face Transformers
- Databases: PostgreSQL, Redis
- Cloud: AWS
- Containers: Docker
- CI/CD: GitHub Actions
- Version control: Git
- Collaboration: Slack, Jira, Notion
- AI tools: ChatGPT, Claude, Cursor
Before you apply
- Any false information in your application, or found during the process, may end your candidacy
- Candidates eligible for employment protection are given preference under applicable law
TTS AI Research Engineer
Full-timeAlternative military service eligible
Hiring process
- Document screening
- Résumé / portfolio in any format, as a PDF
- A detailed description of your TTS projects and research is required
- Email it to recruit@mindlogic.ai
- Interviews
- 1. Online technical interview (including a portfolio presentation)
- 2. In-person Tech & Culture Fit interview
What you’ll do
The core problems we’re solving now
- Natural intonation: TTS models that accurately reproduce each persona’s way of speaking and intonation
- Richer emotional expression: TTS that can express emotion naturally — laughter, sighs, interjections
- Minimal first-token latency: an ultra-low-latency TTS system for real-time conversation
Core R&D areas
- Realtime Conversational Voice Cloning: real-time voice cloning tuned for conversation
- Persona-based Expressive TTS: a personalized speech synthesis engine that reflects each character’s own voice style and emotion
- Neural Audio Codec optimization
- TTS data pipeline: improving model performance through speech data preprocessing, cleaning and augmentation
Requirements
- Master’s degree or higher in computer science, electrical engineering or a related field, or equivalent hands-on experience
- 3+ years of TTS / speech synthesis R&D
- Deep understanding of, and hands-on experience with, deep learning frameworks such as PyTorch and TensorFlow
- Have implemented and customized modern deep-learning TTS algorithms (FastSpeech, VITS, XTTS, etc.)
- Speech signal processing fundamentals: understanding and use of FFT, STFT, mel-spectrograms, MFCCs and the like
- Have built TTS training pipelines end to end: data preprocessing, training, inference and tuning
- Proficiency in Python and audio processing libraries (librosa, torchaudio, etc.)
- English good enough to read and write technical documents
Nice to have
- Have implemented real-time conversational TTS (especially minimizing first-token latency)
- Emotional & Expressive TTS: have implemented natural laughter, sighs and interjections
- Experience developing voice cloning and conversational TTS
- Speech synthesis papers at international conferences: Interspeech, ICASSP, NeurIPS, ICLR, etc.
- Have experimented with recent TTS models: VITS, XTTS, NeuralSpeech, SpeechT5, Bark, CSM, etc.
- Neural vocoder optimization: real-time inference with WaveNet, WaveGlow, HiFi-GAN, BigVGAN, etc.
- Have experimented with and optimized neural audio codec models: SNAC, SoundStream, EnCodec, etc.
- Background in phonetics or linguistics (understanding intonation patterns)
- Have applied and operated TTS in commercial services (API servers, deployment, etc.)
- MLOps and model serving experience (Docker, Kubernetes, cloud services)
Tech stack
- Languages: Python, TypeScript/JavaScript
- Deep learning frameworks: PyTorch, TensorFlow, Hugging Face Transformers
- Databases: PostgreSQL, Redis
- Cloud: AWS
- Containers: Docker
- CI/CD: GitHub Actions
- Version control: Git
- Collaboration: Slack, Jira, Notion
- AI tools: ChatGPT, Claude, Cursor
Before you apply
- Any false information in your application, or found during the process, may end your candidacy
- Candidates eligible for employment protection are given preference under applicable law
Contact: recruit@mindlogic.ai
Deadline: rolling — we may close early once we find the right person
