Job Summary
Experience:
15.00 - 25.00 Years
Industrial Type:
IT-Software/Software Services
Location:
Bangalore
Functional Area:
IT Software - Other
Designation:
Principal Architect - AI
Key Skills:
Principal Architect
Educational Level:
Graduate/Bachelors
Job Post Date:
2026-06-24 16:09:34
Stream of Study:
Degree:
BE-Comp/IT, BE-Other, BTech-Comp/IT, BTech-Other, MBA, MCA, MTech-Comp/IT, MTech-Other
Company Description
The client is a US headquartered company founded in 2013 and is an award winning applied AI and data science software and services company. They specialise in doing work in Big Data, Decision Science, Predictive Analytics, Machine Learning, Consulting, AI, Data Analytics, Deep Learning, Generative AI, LLMs, and AI Agent.
Job Description
Define the stages of the LLM lifecycle: Continuous Pre-training (CPT) on domain-specific data, SFT, and alignment (DPO/PPO/RLHF).
? Design the infrastructure for multi-node, multi-GPU training. You will make the call on using the right infrastructure
? Own the data strategy. You will decide the fine tuning between general-purpose data and domain-specific corpora to prevent catastrophic forgetting.
? Manage infrastructure optimize training efficiency (TFLOPS/GPU) and implement checkpointing strategies to ensure zero loss of progress during hardware failures.
? Build a comprehensive, sovereign evaluation suite (beyond standard benchmarks) that measures the model`s performance on the client’s specific data.
? Act as the primary technical point of contact for the customer’s team, translating their requirements into training approach and solution design.
? Expertise in mitigating loss spikes, adaptive gradient clipping, and sophisticated LR schedules
? Hands-on experience with Mixed Precision (BF16/FP16) and FP8 workflows, including dynamic loss scaling management
? Leveraging Grokking techniques and monitoring gradient health to adjust curriculum mid-run.
? Implementing curriculum learning and domain-specific mixture ratios to prevent catastrophic forgetting during Continuous Pre-training (CPT).
