We are looking for a results-oriented Applied AI Research Engineer who thrives on adapting pre-trained models to real product problems across multiple domains.
Key Responsibilities of the Team
- Build AI-Powered Features: Implement product features across ASR, NLP domains powered by pre-trained foundation and task-specific models.
- Adapt Pre-Trained Models: Match, fine-tune, and orchestrate pre-trained models (ASR, acoustic, LLMs) to specific challenges.
- Prompt & Adaptation Engineering: Design, test, and iterate prompts, fine-tuning recipes, and adaptation strategies to achieve desired behaviors from language, and speech models.
- RAG & Agentic Workflows: Implement retrieval-based pipelines and chain-of-thought/agentic reasoning flows.
- Evaluation Pipelines: Run quantitative and qualitative evaluations (using automation and human feedback) to measure performance, correctness, and UX.
- Data Preparation: Work with raw data to structure datasets needed for inference and evaluation.
- Close Collaboration with Product Team: Translate research strategies into scalable, production-ready applications.
- Production Integration: Collaborate with backend/frontend engineers to deploy and monitor AI features in real-world environments.
- AI Pipelines: Build and maintain end-to-end pipelines that combine speech, and language components into reliable production workflows.
What You’ll Do
- Own the full adaptation lifecycle: understand product needs and technical context, survey existing pre-trained models and techniques, prototype, evaluate, iterate, and deploy.
- Work across ASR and acoustic modeling, NLP (generative and classical), applying the right pre-trained tool to each problem rather than building from scratch.
- Partner with dynamic, cross-functional squads across engineering, research, product, and operations.
- Apply practical, outcome-focused research methodologies, shaping solutions that directly serve product goals and customer needs.
- Design evaluation frameworks and meaningful metrics that guide experimentation and drive continuous improvement.
- Write production-ready Python code, integrating algorithms and models directly into Verbit’s platform.
- Prioritize effectively using Pareto thinking, delivering high-value quick wins while supporting long-term strategic advancements.
- Serve as a key technical member whose ideas, perspective, and ownership significantly influence the team’s direction.
What You Bring
- Experience with data-driven applied AI work across one or more domains (speech, NLP, vision), including evaluation design and metric-driven experimentation.
- 3+ years of experience in applied AI or ML engineering, preferably 5+.
- Hands-on expertise adapting pre-trained models — LLMs, ASR/acoustic models, models — in applied, product-oriented settings. Comfort matching existing techniques to problems rather than designing new algorithms.
- Excellent Python engineering skills, with a strong track record translating prototypes into scalable ML or algorithmic solutions.
- Strong understanding of model behavior, ML development workflows, experimentation processes, and fine-tuning / adaptation strategies.
- Ability to work effectively in fast-paced, iterative environments and collaborate closely with engineering teams.
- Fluency with AI coding agents and tools, and a habit of continuously adopting new ones — while still owning the engineering logic and the code you ship.
Who You Are
- Positive, energetic, and proactive, with growth mindset and a strong sense of ownership.
- A clear communicator who can make technical reasoning accessible across functions.
- Motivated by creating real value for users and the business, and excited to contribute where impact is highest.
- Flexible and comfortable adjusting priorities based on product needs and feedback.
- A team-oriented contributor who thrives in dynamic, cross-disciplinary squads.
- Eager to leverage AI coding assistants to move faster, but rigorous about reviewing, understanding, and taking responsibility for every line that goes in.
Nice to Have
- Experience with real-time or streaming systems.
- Background in speech processing, acoustic modeling, conversation intelligence, or multimodal AI pipelines.
- Infrastructure knowledge (e.g., AWS), and familiarity with GitHub and Linux.
