Senior Software Engineer - Agentic AI Applications

Job Description

Position: Senior Software Engineer — Agentic AI Applications

Location: India — Remote or Hybrid (Delhi/NCR office available)

Experience: 7+ years

About Chegg Skills At Chegg Skills, our software empowers learners to move from learning to earning. We build applications and systems that help motivated career switchers make life-changing transitions into high-growth roles. We take pride in making education accessible, and we build the technology that helps learners achieve mastery and success at scale. You'll join a high-ownership team of motivated engineers rethinking what modern education can look like. We serve thousands of learners and educators each year, and your work will directly shape how they learn and succeed.

About The Role We are hiring a Senior Software Engineer to design, build, and own production-grade agentic AI applications that sit at the core of the Chegg Skills learning experience. You will work across the full AI stack: multi-agent orchestration, retrieval systems, evaluation pipelines, LLM application engineering, learner-facing experiences, and the operator tools that make all of it possible. This role is

full-stack with AI depth : you will lead backend systems and AI engineering while also shipping the React interfaces that learners and operators actually use. You will move between model experimentation, production engineering, and product iteration — building systems that are reliable, evaluable, and cost-efficient at scale. This is a long-arc role. The team is building the first version of the Agentic Skills Framework, but the platform will keep evolving for years. We are hiring for the engineer who will shape what comes next, not just ship the first release.

Tech Stack

• Backend: Python (FastAPI) primary; Kotlin/Java (Spring Boot) in adjacent services; GraphQL (DGS/Apollo); REST
• AI / Models: OpenAI (GPT-4 family), Anthropic Claude (via AWS Bedrock), Google Gemini (via Vertex AI), Meta Llama — accessed through a centralized Databricks AI Gateway with routing, fallbacks, and cost/latency controls
• Agentic & RAG: LangChain (ReAct agents, custom tools), territory-restriction patterns, multi-collection vector search (Zilliz/Milvus), OpenAI embeddings, hybrid/lexical retrieval
• AI Safety: AWS Bedrock Guardrails, custom prompt-injection defenses, PII filtering, content moderation
• Workflow: Temporal for async pipelines, retries, and long-running AI tasks
• Storage & Cache: DynamoDB, PostgreSQL, Redis (ElastiCache)
• Frontend: React + TypeScript, SWR, component libraries for AI surfaces (chat, lesson rendering, dashboards)
• Auth & Platform: Auth0 (JWT + JWKS), AWS (ECS, Lambda, Secrets Manager), Docker, GitLab CI
• Observability: New Relic, Rootly

What You Will Do

• Architect and ship production agentic AI applications — multi-step workflows, tool use, structured outputs, territory-restricted reasoning, and confidence-scored validation loops.
• Design and improve retrieval systems across multiple modes: catalog-level retrieval for curriculum matching, content-level RAG for in-lesson Q&A, and lexical or graph approaches where vector search underperforms.
• Build the LLM application layer: prompt orchestration, structured outputs, function/tool calling, model routing via the AI gateway, conversation memory, and context management.
• Ship learner-facing AI experiences (tutoring, practice, evaluation, coach) and operator-facing tools (program authoring, evaluation dashboards, admin) end-to-end in React + TypeScript.
• Build evaluation infrastructure that makes AI quality measurable — gold datasets, rubrics, regression pipelines, threshold calibration, factuality and groundedness metrics, A/B test scaffolding.
• Implement AI safety: prompt-injection defenses, PII detection and redaction, content moderation, abuse detection, and policy enforcement on learner-facing surfaces.
• Own production reliability for AI systems: model failover, rate limiting, timeout and retry policies, cost monitoring, observability, and incident response.
• Improve answer quality, groundedness, latency, and cost through better prompts, retrieval, model selection, data quality, and evaluation loops.
• Partner with product, design, data engineering, learning design, and platform/SRE teams to deliver reliable AI features with measurable learner impact.
• Provide technical mentorship and help shape engineering standards, architecture, and platform direction as the product scales.

What We Are Looking For

• 7+ years of professional software engineering experience, with strong full-stack depth across backend services and product surfaces.
• Production experience with Python (FastAPI/Flask) for building scalable AI services and APIs; familiarity with Kotlin/Java is a plus.
• Production experience with React + TypeScript building real product surfaces — learner-facing AI experiences (chat, content rendering, dashboards) and operator/admin tools.
• Hands-on production experience building LLM applications — prompt engineering, structured outputs (JSON schemas, function/tool calling), conversation memory, and evaluation.
• Practical experience with agentic AI patterns: multi-step orchestration, ReAct or planner-style agents, custom tool integration, iteration limits, and confidence scoring.
• Production RAG experience end-to-end: chunking strategies, embeddings, vector databases (Milvus/Zilliz, Pinecone, pgvector, or similar), threshold calibration, hybrid/lexical retrieval, grounding, and hallucination mitigation.
• Experience designing AI product surfaces that handle streaming responses, partial outputs, loading states, tool-call visualization, and recoverable errors — with care for learner experience.
• Experience working across multiple LLM providers (OpenAI, Anthropic, Google) with a clear sense of cost, latency, quality, and context trade-offs.
• AI safety and guardrails experience: prompt-injection defenses, PII detection/redaction, content moderation, and evaluation of unsafe outputs.
• AI evaluation discipline: gold datasets, rubrics, human-labeled and synthetic test sets, automated regression evaluation, and threshold tuning against curated test data.
• Strong production engineering fundamentals — API design, observability, security, performance, cache invalidation, and async/streaming — plus AI-specific reliability: fallback chains, rate limiting, cost/latency monitoring, and incident response.
• Experience designing systems that handle consumer PII responsibly — data minimization, access controls, retention policies, and awareness of regulatory requirements for consumer-data protection.

Preferred Qualifications

• Experience shipping AI features for education, assessments, tutoring, grading, or personalized learning.
• Experience with at least one AI gateway (Databricks AI Gateway, Portkey, LiteLLM, or similar) for model routing, fallbacks, rate limiting, and cost/observability.
• Production experience with at least one agentic/RAG framework — LangChain (especially ReAct + custom tools), LlamaIndex, or DSPy.
• Experience with at least one AI evaluation/observability platform — RAGAS, LangSmith, Arize Phoenix, DeepEval, or similar.
• Experience with workflow orchestration (Temporal, Airflow, or similar) for async pipelines and long-running AI tasks.
• Experience with the production platform stack — Auth0 (JWT/JWKS), AWS at scale (ECS, Lambda, Bedrock, Secrets Manager), Docker, GitLab CI — plus CORS and session security for multi-tenant consumer apps.

Impact You Can Expect to Have

• Cut time-to-ship for new AI features by improving agentic/RAG platform foundations and developer ergonomics.
• Improve learner outcomes on practice, assessment, and skill-attainment workflows through higher-quality AI support.
• Raise the engineering bar for agentic AI, retrieval, evaluation, and production reliability across the team and the org.

Work Style

• High ownership of technical design, delivery, and production outcomes.
• Fast iteration with strong engineering discipline — testing, observability, documentation, and design reviews.
• Close collaboration with product and design to turn ambiguous AI opportunities into shipped capabilities.
• Practical use of AI coding tools to move faster without compromising maintainability, security, or quality. Why do we exist? Students are working harder than ever before to stabilize their future. Our recent research study called State of the Student shows that nearly 3 out of 4 students are working to support themselves through college and 1 in 3 students feel pressure to spend more than they can afford. We founded our business on provided affordable textbook rental options to address these issues. Since then, we’ve expanded our offerings to supplement many facets of higher educational learning through Chegg Study, Chegg Math, Chegg Writing, Chegg Internships, Chegg Skills, and more to support students beyond their college experience. These offerings lower financial concerns for students by modernizing their learning experience. We exist so students everywhere have a smarter, faster, more affordable way to student.

Video Shorts Life at Chegg: http://youtu.be/Fwf90zgaOLA Chegg Corporate Career Page: https://jobs.chegg.com/ Chegg India: http://www.cheggindia.com/ Chegg Israel: http://www.chegg.com/about/working\-at\-chegg/israel/ Chegg Skills: https://www.chegg.com/skills

Chegg out our culture and benefits! http://www.chegg.com/about/working\-at\-chegg/benefits/ Chegg is an equal opportunity employer