Lead AI Architecture

Sahana System Limited; Ahmedabad District; On-site; INR 3,500,000 - 5,200,000; Full time

Benefits offered by this job

Access to large GPU clusters

Leadership role with board visibility

Job summary

Sahana System Limited is seeking a visionary AI Architect to own end-to-end architectures for next-generation Sovereign AI platforms. You will drive strategy for CV, GenAI, LLMs/SLMs, RAG, and multi-modal systems deployed across defence, enterprise, and public sectors.

You will lead a 20+ person AI/ML team, set technical standards, and shape governance, data residency, and edge optimizations for multi-year programs.

Qualifications

• 15+ years hands-on in AI/ML with end-to-end project implementation across CV, GenAI, RL.
• At least 5 years in architectural leadership roles.
• Experience with large-scale public sector/defence projects is preferred.

Responsibilities

• Technical leadership to architect and lead large-scale AI systems.
• Design and scale proprietary LLMs/SLMs, RAG frameworks, and domain-specific Sovereign AI models.
• Lead indigenous, secure, on-prem AI stacks with data residency and auditability.
• Build and mentor 20+ AI/ML engineers and researchers; set standards.
• Own architecture for multi-year programs spanning data pipelines to model serving at scale.
• Collaborate with product, security, and business teams to translate requirements.
• Evaluate SOTA research, publish patents, and define AI governance.

Skills

PyTorch/TensorFlow

CUDA

Transformer architectures

Diffusion models

YOLO

NeRFs

3D vision

RAG design

MLOps

LMMs/SLMs

Education

M.Tech/PhD in CS/AI/ECE or related

Tools

Kubeflow

MLflow

Ray

Triton

vLLM

Job description

Position Overview: We are seeking a visionary AI Architect with 15+ years of deep, hands-on expertise to define and own the end-to-end architecture of our next-generation AI platforms. This role will drive strategy and technical execution for large-scale Sovereign AI initiatives spanning Computer Vision, Generative AI, LLMs/SLMs, RAG, and multi-modal systems deployed across defence, enterprise, and public sector.

Key Responsibilities

1)Technical Leadership: Architect and lead development of large-scale AI systems covering:

• Generative AI: Text, audio, and video generation; diffusion models, GANs, transformers; multi-modal synthesis
• Core ML: Deep Learning, Reinforcement Learning, Active Learning pipelines from research to production

2)LLM/SLM & RAG Platforms: Design and scale proprietary LLMs/SLMs, RAG frameworks, vector DBs, agentic workflows, and fine-tuning pipelines for domain-specific Sovereign AI models.

3)Sovereign AI: Lead initiatives for indigenous, secure, on-prem AI stacks with data residency, compliance, and model auditability for strategic sectors.

4)Team Leadership: Build, mentor, and manage a team of 20+ AI/ML engineers, research scientists, and MLOps specialists. Set technical standards, review architectures, and drive engineering excellence.

5)Project Delivery: Own architecture for complex, multi-year programs involving data pipelines, training infra, GPU clusters, model serving at scale, and edge optimization

6)Cross-functional Collaboration: Partner with product, data engineering, security, and business teams to translate mission requirements into scalable AI solutions

7)Innovation & Governance: Evaluate SOTA research, publish patents, define AI governance, model risk, and ethical AI frameworks.

Required Qualifications & Skills

1)15+ years hands-on in AI/ML with proven end-to-end project implementation across Computer Vision, GenAI, and RL. At least 5 years in architectural leadership roles.

Technical Depth:

• Expert in PyTorch/TensorFlow, CUDA, transformer architectures, diffusion models, YOLO/SAM, NeRFs, 3D vision
• Production experience with LLMs/SLMs: pre-training, PEFT/LoRA/QLoRA, quantization, distributed training on 100+ GPUs
• Strong RAG design: embeddings, rerankers, hybrid search, knowledge graphs, evaluation frameworks
• RLHF, Active Learning, continual learning, and MLOps: Kubeflow, MLflow, Ray, Triton, vLLM Deployed models at scale: >10M inferences/day, latency <100ms, edge devices Leadership: Managed and scaled teams of 20+ AI engineers. Built hiring pipelines, OKRs, and technical career ladders

• 2)Domain: Experience with large-scale public sector/defence projects preferred.
• 3)Understanding of data sovereignty, security clearances, and air-gapped deployments is a plus

Must-have:

• M.Tech/PhD in CS, AI, ECE or related field from premier institute preferred.
• 15+ years of hands-on experience

Skills & Attributes

• Experience building foundation models from scratch
• Contributions to open-source AI projects or top-tier conferences
• Prior work with MoD, DPSUs, or critical infrastructure AI deployments
• Integrity & strong communication

What We Offer

• Opportunity to architect India’s Sovereign AI stack
• Access to large GPU clusters and petabyte-scale datasets
• Leadership role with board visibility