Speech Research Intern-2

Centific

2d•Hybrid

About The Position

Centific AI Research seeks a PhD Research Intern to design and evaluate speech-first models, with a focus on Spoken Language Models (SLMs) that reason over audio and interact conversationally. You’ll move ideas from prototype to practical demos, working with scientists and engineers to deliver measurable impact.

Requirements

PhD candidate in CS/EE (or related) with research in speech, audio ML, or multimodal LMs.
Fluency in Python and PyTorch, with hands-on GPU training.
Familiarity with torchaudio or librosa.
Working knowledge of modern sequence models (Transformers or SSMs) and training best practices.
Depth in at least one area: (a) discrete speech tokens/temporal compression, (b) modality alignment to LLMs via adapters, or (c) post-training/instruction tuning for speech tasks.
Strong experimentation habits: clean code, ablations, reproducibility, and clear reporting.

Nice To Haves

Experience with speech generation (neural codecs/vocoders) or hybrid text+speech decoding.
Background in multilingual or code-switching speech and domain adaptation.
Hands-on work evaluating safety, bias, hallucination, or spoofing risks in speech systems.
Experience with distributed training/serving (FSDP/DeepSpeed).
Experience with ESPnet, SpeechBrain, or NVIDIA NeMo.

Responsibilities

Design and evaluate speech-first models, with a focus on Spoken Language Models (SLMs) that reason over audio and interact conversationally.
Move ideas from prototype to practical demos, working with scientists and engineers to deliver measurable impact.
Develop end-to-end speech dialogue systems (speech-in/speech-out) and speech-aware LLMs.
Align speech encoders and text backbones via lightweight adapters.
Implement efficient speech tokenization and temporal compression suitable for long-form audio.
Conduct reliable evaluation across recognition, understanding, and generation tasks, including robustness and safety.
Optimize for latency-aware inference for streaming and real-time user experiences.
Prototype a conversational SLM using an SSL speech encoder and a compact adapter on an existing LLM.
Create a data recipe that blends conversational speech with instruction-following corpora.
Build an evaluation harness that covers ASR/ST/SLU and speech QA, including streaming metrics (latency, stability, endpointing).
Ship a minimal demo with streaming inference and logging.
Author a crisp internal write-up: goals, design choices, results, and next steps for productionization.