Software Engineering SMTS- LLM Model building
Salesforce · Bengaluru · Posted 2026-07-06
Tech stack: Kubernetes, Python
Apply on the company site · Get a referral for this role
Salesforce salary & ratings · Salesforce interview process · More live openings
About the role
Senior Applied Scientist – AgentForce Team Overview The AgentForce Data Science team powers the core Large Language Models (LLMs) behind Salesforce’s production-grade AI agents. Our work directly impacts millions of users by enabling trustworthy, scalable, and high-performance AI systems across customer support, sales, marketing, analytics, and internal productivity workflows. We operate at the intersection of cutting-edge research and real-world deployment, owning the full model development lifecycle—from research ideation and training to fine-tuning, evaluation, continuous learning, and production rollout. Role Overview We are seeking a strong Senior Applied Scientist to contribute to advanced LLM research and model development for AgentForce’s production AI services. This role requires strong hands-on involvement across the full model development lifecycle, including model training, fine-tuning, evaluation, reinforcement learning, optimization, and deployment support. The ideal candidate is a strong individual contributor who can independently drive technical execution while collaborating closely with research, engineering, product, and infrastructure teams. The candidate will work on production-scale AI systems supporting enterprise-grade agentic workflows, reasoning systems, evaluation services, and multi-modal AI capabilities. Key Responsibilities Research, Modeling & Hands-On Execution Execute hands-on work across the full model development lifecycle, including: Data preparation and curation Synthetic data generation Model training and fine-tuning RLHF / RLAIF workflows Evaluation and benchmarking Error analysis and iteration Inference optimization Deployment readiness Contribute to research and development efforts for: Large language models Tool-calling systems Agentic reasoning workflows Multi-modal AI models Evaluation and guardrails systems Continuous learning pipelines Design and implement experimentation pipelines for: Reinforcement learning Preference optimization Alignment tuning Offline and online feedback learning Conduct rigorous experimentation, benchmarking, and failure analysis to improve: Accuracy Latency Reliability Robustness Cost efficiency Translate research ideas into scalable production-ready AI solutions. Support optimization initiatives including: Quantization Distillation Distributed inference optimization Throughput and serving efficiency improvements Technical Collaboration Partner with senior scientists, engineers, and product teams to deliver production AI solutions. Contribute to model training, evaluation, release readiness, and production support processes. Collaborate with infrastructure teams on scalable training and inference workflows. Help define and improve best practices for: Model evaluation Experiment tracking Data quality Continuous learning Production monitoring Participate in technical reviews, roadmap discussions, and cross-functional planning efforts. Mentorship & Growth Mentor junior team members through technical guidance and collaboration. Contribute to a strong culture of: Scientific rigor Ownership Reproducibility Fast iteration Operational excellence Stay current with advancements in: LLMs Reinforcement learning Agentic AI Multi-modal AI Distributed AI systems Contribute to internal technical knowledge sharing and innovation initiatives. Required Qualifications Education & Research Background PhD or Master’s degree in Computer Science, Machine Learning, Artificial Intelligence, or a related field. Strong research or industry experience in areas such as: LLMs NLP Reinforcement learning Multi-modal AI Agentic systems Core Technical & Hands-On Requirements Demonstrated hands-on experience in model development, training, fine-tuning, evaluation, and experimentation. Strong expertise in: Large language model fine-tuning Model evaluation Inference optimization Continuous learning workflows Experience with: Reinforcement learning Preference learning Human-in-the-loop systems Production AI evaluation Understanding of: AI safety Guardrails Reliability Production AI systems Experience working with distributed training or large-scale inference systems. Coding & Tooling Strong proficiency in Python with solid software engineering fundamentals. Experience with: PyTorch TensorFlow Familiarity with modern LLM tooling and infrastructure, including: Hugging Face (Transformers, PEFT, Accelerate) DeepSpeed FSDP Ray Kubernetes vLLM Strong experimentation and data analysis skills using: NumPy Pandas Custom evaluation pipelines Leadership & Collaboration Strong collaboration and communication skills across: Research Engineering Product Infrastructure teams Ability to independently drive technical projects and deliver high-quality execution. Comfortable working in fast-moving, highly iterative AI development environments. Preferred Qualifications Experience deploying and supporting production AI systems at scale. Background in: Enterprise AI systems Agentic AI workflows Tool-calling systems Multi-agent systems Familiarity with: AI trust and safety systems Governance frameworks Responsible AI practices Experience with: Multi-modal AI Long-context models Retrieval-augmented systems Planner and reasoning systems Experience with multi-GPU or distributed compute environments. Why Join AgentForce? Work on mission-critical AI systems operating at massive enterprise scale. Build and deploy production-grade LLM systems used by millions of users. Collaborate with world-class researchers and engineers. Solve challenging problems in reasoning, reinforcement learning, multi-modal AI, and agentic systems. See your work directly impact real-world AI products and enterprise customers.
Qualifications
- Experience deploying and supporting production AI systems at scale.
- Background in: Enterprise AI systems Agentic AI workflows Tool-calling systems Multi-agent systems
- Familiarity with: AI trust and safety systems Governance frameworks Responsible AI practices
- Experience with: Multi-modal AI Long-context models Retrieval-augmented systems Planner and reasoning systems
- Experience with multi-GPU or distributed compute environments.
- Why Join AgentForce?
- Work on mission-critical AI systems operating at massive enterprise scale.
- Build and deploy production-grade LLM systems used by millions of users.
- Collaborate with world-class researchers and engineers.
- Solve challenging problems in reasoning, reinforcement learning, multi-modal AI, and agentic systems.
- See your work directly impact real-world AI products and enterprise customers.
Responsibilities
- Research, Modeling & Hands-On Execution Execute hands-on work across the full model development lifecycle, including: Data preparation and curation Synthetic data generation Model training and fine-tuning RLHF / RLAIF workflows Evaluation and benchmarking Error analysis and iteration Inference optimization Deployment readiness
- Contribute to research and development efforts for: Large language models Tool-calling systems Agentic reasoning workflows Multi-modal AI models Evaluation and guardrails systems Continuous learning pipelines
- Design and implement experimentation pipelines for: Reinforcement learning Preference optimization Alignment tuning Offline and online feedback learning Conduct rigorous experimentation, benchmarking, and failure analysis to improve: Accuracy Latency Reliability Robustness Cost efficiency Translate research ideas into scalable production-ready AI solutions.
- Support optimization initiatives including: Quantization Distillation Distributed inference optimization Throughput and serving efficiency improvements Technical Collaboration
- Partner with senior scientists, engineers, and product teams to deliver production AI solutions.
- Contribute to model training, evaluation, release readiness, and production support processes.
- Collaborate with infrastructure teams on scalable training and inference workflows.
- Help define and improve best practices for: Model evaluation Experiment tracking Data quality Continuous learning Production monitoring
- Participate in technical reviews, roadmap discussions, and cross-functional planning efforts.
- Mentorship & Growth Mentor junior team members through technical guidance and collaboration.
- Contribute to a strong culture of: Scientific rigor Ownership Reproducibility Fast iteration Operational excellence
- Stay current with advancements in: LLMs Reinforcement learning Agentic AI Multi-modal AI Distributed AI systems
- Contribute to internal technical knowledge sharing and innovation initiatives.
- Education & Research Background PhD or Master’s degree in Computer Science, Machine Learning, Artificial Intelligence, or a related field.
- Strong research or industry experience in areas such as: LLMs NLP Reinforcement learning Multi-modal AI Agentic systems Core Technical & Hands-On Requirements
- Demonstrated hands-on experience in model development, training, fine-tuning, evaluation, and experimentation.
- Strong expertise in: Large language model fine-tuning Model evaluation Inference optimization Continuous learning workflows
- Experience with: Reinforcement learning Preference learning Human-in-the-loop systems Production AI evaluation
- Understanding of: AI safety Guardrails Reliability Production AI systems
- Experience working with distributed training or large-scale inference systems.
- Coding & Tooling Strong proficiency in Python with solid software engineering fundamentals.
- Experience with: PyTorch TensorFlow
- Familiarity with modern LLM tooling and infrastructure, including: Hugging Face (Transformers, PEFT, Accelerate) DeepSpeed FSDP Ray Kubernetes vLLM
- Strong experimentation and data analysis skills using: NumPy Pandas Custom evaluation pipelines Leadership & Collaboration
- Strong collaboration and communication skills across: Research Engineering Product Infrastructure teams
- Ability to independently drive technical projects and deliver high-quality execution.
- Comfortable working in fast-moving, highly iterative AI development environments.