Flag job

Report

Lead Machine Learning Engineer - Agentic AI

Min Experience

6 years

Location

san francisco, hyderabad

JobType

full-time

About the job

Info This job is sourced from a job board

About the role

Job Title: Lead Machine Learning Engineer – Agentic AI & Data Quality Location: San Francisco, CA, US & Hyderabad, Telangana, India Experience: 6+ years About Us At Deccan AI, founded by alumni of IIT Bombay and IIM Ahmedabad, we specialize in LLM model development and AI-first scaled operations. With headquarters in San Francisco and an advanced delivery hub in Hyderabad, our mission is to build AI for Good — creating technology that drives innovation with meaningful societal impact. About the Role We are seeking a Lead Machine Learning Engineer with 6+ years of experience to spearhead our Agentic AI initiatives and establish enterprise-grade data quality standards across our ML systems. This is a high-impact leadership role where you'll architect autonomous agent frameworks, build production-scale LLM systems, and ensure our training datasets meet the highest standards of accuracy, safety, and relevance for enterprise deployments. You'll lead the end-to-end ML lifecycle — from designing agentic workflows and multi-agent orchestration to implementing automated QA pipelines, evaluation strategies, and production monitoring systems. As the technical leader and strategic advisor, you'll bridge engineering, research, and client success teams while mentoring junior engineers and driving architectural decisions for enterprise AI solutions. This role demands both deep technical expertise in agentic AI systems and the ability to translate complex ML concepts into scalable, production-ready solutions that meet Fortune 500 client expectations. Key Responsibilities Agentic AI Leadership: Design and architect multi-agent systems using frameworks like LangChain, AutoGen, CrewAI, or custom orchestration layers Build autonomous agents with tool-use capabilities, memory systems, and chain-of-thought reasoning Implement agent planning algorithms (ReAct, Tree-of-Thoughts, Plan-and-Execute patterns) Lead research and development of self-improving agent systems with feedback loops Design agent evaluation frameworks measuring task success, efficiency, and safety Enterprise-Scale Data Quality & MLOps: Architect automated QA pipelines for SFT transcripts, RLHF preference pairs, and multi-turn conversational data at enterprise scale Implement real-time data validation systems with schema enforcement, semantic checks, and embedding-based deduplication Build production-grade safety filters detecting toxicity, bias, PII leakage, and adversarial inputs Establish data versioning, lineage tracking, and audit trails for regulatory compliance (GDPR, SOC 2) Design monitoring dashboards tracking data drift, distribution shifts, and quality metrics LLM Fine-Tuning & Model Development: Lead fine-tuning initiatives using advanced techniques (LoRA, QLoRA, full fine-tuning) on enterprise-scale datasets Architect reward modeling pipelines for RLHF/RLAIF with multi-objective optimization Implement distributed training strategies across GPU clusters for 7B-70B+ parameter models Build custom evaluation harnesses combining human raters, LLM-as-judge, and domain benchmarks Optimize inference pipelines for latency, throughput, and cost at production scale Evaluation & Benchmarking Strategy: Design comprehensive evaluation frameworks covering accuracy, safety, hallucination detection, and alignment Build proprietary benchmarks tailored to client domains (legal, finance, healthcare, etc.) Implement statistical rigor in evaluations (confidence intervals, significance testing, inter-rater reliability) Develop automated critique generation systems for model output analysis Create client-facing evaluation dashboards with interpretable metrics Team & Stakeholder Leadership: Mentor and grow a team of ML engineers and data scientists Partner with annotation teams to design labeling guidelines and resolve edge cases Drive technical roadmap aligned with product and business objectives Present technical insights to C-suite executives and enterprise clients Collaborate cross-functionally with platform engineering, research, and sales teams Research & Innovation: Prototype novel architectures for agent reasoning, tool integration, and multi-modal capabilities Develop proprietary metrics for signal quality (reward entropy, preference consistency, data difficulty) Publish findings and contribute to open-source ML tooling Stay current with latest research in agentic AI, LLMs, and alignment techniques Impact: As Lead MLE, you'll set the technical standard for how Deccan AI builds enterprise-grade agentic systems. Your architectural decisions will influence millions of AI interactions, your quality frameworks will ensure client trust, and your leadership will shape the next generation of ML talent. You'll be the bridge between cutting-edge research and production systems that create real business value.

About the company

At Deccan AI, founded by alumni of IIT Bombay and IIM Ahmedabad, we specialize in LLM model development and AI-first scaled operations. With headquarters in San Francisco and an advanced delivery hub in Hyderabad, our mission is to build AI for Good — creating technology that drives innovation with meaningful societal impact.

Skills

python
pytorch
hugging face
langchain
llama index
autogen
semantic kernel
docker
kubernetes
airflow
dagster
ci/cd
vector databases
pinecone
weaviate
qdrant
sql
ray
spark
aws
gcp
azure