Post-Training of LLM: Innovation in Life Sciences AI

Push the Boundaries of LLM Performance – Lead Post-Training Innovation in Life Sciences AI

At Dizzaroo Pvt Ltd, we are building AI-powered platforms that transform the pharmaceutical and life sciences industry — from drug discovery to digital pathology to clinical trial automation.

Large Language Models (LLMs) are at the core of several of our applications. But we’re not simply plugging in APIs — we’re customizing, aligning, and post-training these models to handle high-stakes, domain-specific biomedical and regulatory content.

We are looking for a seasoned AI/ML expert who understands the full post-training lifecycle of LLMs — not just running a fine-tuning job in a managed platform, but architecting and executing large-scale post-training projects from dataset curation to deployment.

Position

AI/ML Expert – Post-Training of Large Language Models (LLMs)

Location:

Key Responsibilities

• Lead post-training workflows for domain-specific LLMs, including supervised fine-tuning (SFT), reinforcement learning from human feedback (RLHF), and other alignment methods.
• Design and implement large-scale training pipelines — from data preprocessing to distributed training and evaluation.
• Curate domain-specific datasets for biomedical, regulatory, and clinical contexts.
• Experiment with parameter-efficient fine-tuning techniques (LoRA, QLoRA, adapters) and prompt optimization strategies.
• Benchmark model performance on domain-relevant evaluation suites and iterate to improve accuracy, reliability, and explainability.
• Collaborate with product and domain teams to ensure model outputs meet regulatory-grade precision.

Qualifications

• Proven hands-on experience with post-training of LLMs beyond basic managed-service fine-tuning.
• Proficiency in PyTorch and transformers libraries (Hugging Face, DeepSpeed, Megatron-LM, or similar).
• Experience in distributed training, large-scale GPU/TPU clusters, and optimization for high-parameter models.
• Strong understanding of LLM architecture, tokenization strategies, and memory optimization techniques.
• Experience with biomedical or technical text corpora is a plus.
• Bachelor’s/Master’s/PhD in Computer Science, AI/ML, or related field.

What We Value

• End-to-end ownership — you’ve run a project from dataset to deployed model.
• Curiosity and deep technical engagement with model internals.
• The ability to explain complex ML workflows to both engineers and non-technical stakeholders.
• Comfort working in ambiguous, fast-moving R&D environments.

Why Join Us

• Shape the next generation of domain-specific LLMs for life sciences.
• Work at the cutting edge of AI + healthcare with a highly interdisciplinary team.
• Opportunity to set technical direction for high-impact AI initiatives.

How to Apply

Send the following to dhirajg@dizzaroo.com

1. Cover Letter

   – Describe your most significant LLM post-training project.

2. CV/Resume

   .

3. Links to code, research, or demos

   that showcase your work.