Lead Software Engineer - Machine Learning
Freshworks
Bengaluru, Karnataka, IndiaLEAD
Job Description
As a Lead Software Engineer, you will build autonomous intelligent agents.
Responsibilities
- Organizations everywhere struggle under the crushing costs and complexities of “solutions” that promise to simplify their lives. To create a better experience for their customers and employees. To help them grow. Software is a choice that can make or break a business. Create better or worse experiences. Propel or throttle growth. Business software has become a blocker instead of ways to get work done.
- There’s another option. Freshworks. With a fresh vision for how the world works.
- Fresh vision. Real impact. Come build it with us.
- Productionize Autonomous Systems: Serve as the critical link translating advanced research in LLMs, multi-agent frameworks, and cognitive architectures into highly reliable, product-ready, enterprise-scale implementations.
- Scale Multi-Agent Workflows: Build and deploy robust ML infrastructures, runtime environments, and data pipelines engineered to orchestrate and execute millions of autonomous agent decisions with low latency, high efficiency, and safety guardrails.
- Architect Next-Gen AI Frameworks: Drive technical alignment by designing scalable agentic frameworks from the ground up, establishing best practices for prompt engineering frameworks, memory systems, and tool-use integration across the organization.
- Agentic Framework Implementation: Collaborate with AI Researchers and Data Scientists to translate complex reasoning loops (e.g., ReAct, Reflection, Tree of Thoughts) and experimental algorithms into clean, production-grade code.
- End-to-End Agent Architecture: Design, build, and manage comprehensive agent pipelines, including state management, long/short-term vector memory systems, automated tool-calling integrations, and dynamic feedback loops.
- High-Performance Service Delivery: Develop and deploy extensible, scalable API microservices optimized to minimize latency during heavy token generation, parallel agent execution, and high-concurrency traffic loads.
- Operational Telemetry & Evaluation: Design and implement robust tracking frameworks to monitor agent drift, hallucination rates, cost/token efficiency, and multi-step execution accuracy to ensure systemic reliability.
- Strategic Ecosystem Collaboration: Architect foundational AI primitives from scratch and liaise with cross-product architects to ensure seamless integration of agentic capabilities into existing product ecosystems.
- Prototyping & LLM Benchmarking: Lead Proof of Concept (POC) initiatives utilizing diverse tech stacks, open-source orchestrators (e.g., LangGraph, AutoGen, CrewAI), and custom vector infrastructures to validate optimal solutions for complex business workflows.
- Lifecycle Ownership: Independently own the full lifecycle of feature delivery—from alignment with product teams on agent objectives to final production deployment, guardrail reinforcement, and continuous optimization.
Qualifications
- Production Agentic Engineering: Proven capability in translating raw LLMs/LMMs into reliable, stateful, and autonomous software applications with deterministic guardrails.
- LLMOps & Agent Evaluation: Deep expertise in lifecycle management for generative models, including fine-tuning, retrieval-augmented generation (RAG) optimization, prompt versioning, and routing architectures.
- Distributed State & API Design: Strong distributed systems architecture skills, specifically in managing asynchronous workflows, message queues, and low-latency API microservices required for multi-agent coordination.
- Cognitive Telemetry: Proficiency in establishing monitoring frameworks for tracking agent reasoning paths, tool execution success rates, API cost metrics, and user-intent alignment.
- Technical AI Leadership: Ability to execute rapid prototyping with emerging AI stacks, benchmark foundation models, evaluate vector databases, and drive cross-functional alignment.
- Track Record: A proven history of successfully building, productionizing, and maintaining large-scale Machine Learning solutions and scalable backend architectures.