Overview:
We are looking for a skilled AI/ML Engineer to develop, optimize, and maintain intelligent speech-to-text and medical NLP systems. The ideal candidate will work on automating transcription workflows, improving accuracy, reducing turnaround time, and enhancing the quality of medical documents.
Key Responsibilities
-
Develop and fine-tune speech recognition models for doctor dictations (UK, US, India accents).
-
Build NLP pipelines for medical terminology, abbreviations, and context understanding.
-
Implement error detection, auto-correction, and QA automation in transcription outputs.
-
Integrate with voice services such as AWS Transcribe Medical, Azure Cognitive Speech, or Google Speech-to-Text.
-
Create APIs for seamless integration with existing transcription tools, EMR/EHR systems, and workflow software.
-
Work closely with QA and transcription teams to continuously improve model accuracy using real transcription datasets.
-
Ensure data security and HIPAA/GDPR compliance in model training and storage.
Required Technical Skills
-
Strong knowledge of Python
-
Deep understanding of NLP & Machine Learning
-
Experience in speech recognition frameworks such as:
Whisper / Vosk / DeepSpeech / Kaldi
-
Familiarity with transformer models:
BERT, BioBERT, ClinicalBERT, GPT-based medical NLP
-
Experience with Git for code management
-
Experience with cloud services (any one of the following):
AWS, Azure, or GCP
Preferred Skills (Added Advantage)
-
Medical terminology understanding (ICD-10, drugs, symptoms, CPT codes)
-
Experience with FastAPI or Django for backend services
-
Knowledge of data labeling, speech dataset creation, and model evaluation
-
Exposure to LLM fine-tuning or prompt engineering
Soft Skills
-
Clear communication and analytical thinking
-
Problem-solving with minimal supervision
-
Willingness to experiment and build proof-of-concept models
-
Ability to work closely with cross-functional teams (QA, medical reviewers, transcription operators)
Experience Required
-
1–3 years in AI/ML, NLP, or Speech Recognition
-
(Freshers with strong projects are welcome if they show capability in NLP/speech tasks)
