AI Engineer  

About AutoScript

AutoScript is building next-generation AI systems for legal transcription and deposition workflows. We combine real-time speech recognition, LLM pipelines, and structured data extraction to transform how legal teams work with audio and transcripts.

Team

This is a role based in South Jakarta working in a new hybrid office with international colleagues, globally distributed.

What You’ll Work On

You will build systems that process and understand audio at scale, including:

• Analysing audio files using tools like ffmpeg / ffprobe
• Designing pipelines for transcription post-processing (LLMs + rules)
• Extracting structured data (entities, speakers, timestamps)
• Building evaluation and validation systems for AI outputs
• Improving robustness and accuracy of real-world AI systems

This role sits at the intersection of systems engineering + applied AI.

Responsibilities

• Build and maintain audio analysis pipelines (ffmpeg, signal-level insights)
• Design LLM-powered post-processing for transcripts
• Implement structured output extraction (NER, metadata, formatting)
• Define evaluation metrics and validation frameworks
• Work closely with fullstack and QA engineers to productionise systems
• Optimise performance and cost of AI pipelines

Application Process

As part of your application, you will complete a short technical assessment involving:

• audio analysis (ffmpeg / metadata extraction)
• structured output generation
• optional agent-based design

We are looking for builders who can design real systems - not just call APIs.

Requirements

• 2+ years experience in AI / ML / NLP engineering
• Strong Python skills
• Experience working with LLM APIs or open-source models
• Familiarity with embeddings, vector search, or NLP pipelines
• Ability to work with real-world data (messy, noisy inputs)

Nice to Have

• Experience with audio processing or speech systems
• Experience with ffmpeg / ffprobe
• Familiarity with agent frameworks or tool-calling systems
• Experience designing evaluation pipelines for AI systems