I
Innobizvia muse
ML Engineer – LoRA Fine-Tuning | Python, PyTorch & Hugging Face
REMOTEPosted 2w ago
ML EngineerMid LevelFull-time#remote
About the Role
ML Engineer – LoRA Fine\-Tuning \| Python, PyTorch \& Hugging Face
**Experience:** 6\+ Years
**Duration:** 6 Months (Extendable)
**Work Timing:** 11:00 AM – 8:00 PM IST
**Monthly Budget:** ₹1,00,000 – ₹1,11,000
**Work Model:** Offshore Partner Engagement
Job Summary
We are looking for an experienced **Machine Learning Engineer** with strong expertise in LLM fine\-tuning and Parameter\-Efficient Fine\-Tuning (PEFT), particularly LoRA (Low\-Rank Adaptation). The candidate will be responsible for building and operationalizing a LoRA fine\-tuning pipeline to adapt a frozen foundation model for clinical\-domain applications.
The role involves data preprocessing, tokenization, training pipeline implementation, experiment tracking, checkpoint management, adapter packaging, and integration with inference and serving workflows.
Key Responsibilities
* Implement and maintain LoRA fine\-tuning pipelines against frozen foundation models.
* Configure LoRA adapters, training loops, and hyperparameters using PyTorch and Hugging Face PEFT.
* Develop data preprocessing, tokenization, and training/validation splits for clinical datasets.
* Execute, monitor, and optimize training runs, including checkpointing and experiment tracking.
* Maintain model checkpoints, experiment records, and adapter versioning.
* Package trained LoRA adapters as portable artifacts for repeatable deployment.
* Integrate LoRA adapters into model inference and serving pipelines.
* Collaborate with AWS Data Scientists on evaluation frameworks and model\-quality metrics.
* Follow best practices for reproducibility, training efficiency, and model evaluation.
Required Skills
* Strong proficiency in Python and PyTorch.
* Hands\-on experience with Hugging Face Transformers and PEFT.
* Practical experience implementing LoRA and parameter\-efficient fine\-tuning techniques.
* Experience with LLM training, fine\-tuning, and evaluation.
* Knowledge of data preprocessing, tokenization, and dataset preparation.
* Experience with training checkpoints, model versioning, and adapter deployment.
* Familiarity with experiment tracking tools such as MLflow or Weights \& Biases.
* Understanding of model inference and serving workflows.
Preferred Qualifications
* Experience working with clinical or healthcare\-related datasets.
* Familiarity with AWS\-based machine learning workflows.
* Experience optimizing LLM training pipelines and evaluating model performance.
Language Requirement
Candidates must meet the project's mandatory language requirement. Language proficiency will be confirmed during screening.
Engagement Details
* 6\-month contract with the possibility of extension.
* Working hours: 11:00 AM – 8:00 PM IST.
* Offshore partner engagement.
* Monthly rate: ₹1,00,000–₹1,12,000, depending on experience and suitability.
Pay: ₹90,000\.00 \- ₹100,000\.00 per month
Work Location: Remote
**Experience:** 6\+ Years
**Duration:** 6 Months (Extendable)
**Work Timing:** 11:00 AM – 8:00 PM IST
**Monthly Budget:** ₹1,00,000 – ₹1,11,000
**Work Model:** Offshore Partner Engagement
Job Summary
We are looking for an experienced **Machine Learning Engineer** with strong expertise in LLM fine\-tuning and Parameter\-Efficient Fine\-Tuning (PEFT), particularly LoRA (Low\-Rank Adaptation). The candidate will be responsible for building and operationalizing a LoRA fine\-tuning pipeline to adapt a frozen foundation model for clinical\-domain applications.
The role involves data preprocessing, tokenization, training pipeline implementation, experiment tracking, checkpoint management, adapter packaging, and integration with inference and serving workflows.
Key Responsibilities
* Implement and maintain LoRA fine\-tuning pipelines against frozen foundation models.
* Configure LoRA adapters, training loops, and hyperparameters using PyTorch and Hugging Face PEFT.
* Develop data preprocessing, tokenization, and training/validation splits for clinical datasets.
* Execute, monitor, and optimize training runs, including checkpointing and experiment tracking.
* Maintain model checkpoints, experiment records, and adapter versioning.
* Package trained LoRA adapters as portable artifacts for repeatable deployment.
* Integrate LoRA adapters into model inference and serving pipelines.
* Collaborate with AWS Data Scientists on evaluation frameworks and model\-quality metrics.
* Follow best practices for reproducibility, training efficiency, and model evaluation.
Required Skills
* Strong proficiency in Python and PyTorch.
* Hands\-on experience with Hugging Face Transformers and PEFT.
* Practical experience implementing LoRA and parameter\-efficient fine\-tuning techniques.
* Experience with LLM training, fine\-tuning, and evaluation.
* Knowledge of data preprocessing, tokenization, and dataset preparation.
* Experience with training checkpoints, model versioning, and adapter deployment.
* Familiarity with experiment tracking tools such as MLflow or Weights \& Biases.
* Understanding of model inference and serving workflows.
Preferred Qualifications
* Experience working with clinical or healthcare\-related datasets.
* Familiarity with AWS\-based machine learning workflows.
* Experience optimizing LLM training pipelines and evaluating model performance.
Language Requirement
Candidates must meet the project's mandatory language requirement. Language proficiency will be confirmed during screening.
Engagement Details
* 6\-month contract with the possibility of extension.
* Working hours: 11:00 AM – 8:00 PM IST.
* Offshore partner engagement.
* Monthly rate: ₹1,00,000–₹1,12,000, depending on experience and suitability.
Pay: ₹90,000\.00 \- ₹100,000\.00 per month
Work Location: Remote
Similar Jobs
C
Capital One• 1mo ago
Machine Learning Engineer 4 (IC)
McLean, VAFull-time
W
Waystone• 6mo ago
AI Engineer
Leeds, West YorkshireFull-time
A
Aderant• 1mo ago
AI Engineer – Center of Excellence
Atlanta, GAFull-time
S
Signifyd• 1mo ago
Senior Machine Learning Engineer I // II
Chicago, ILFull-time