Headline

AI Engineer | Full Stack Engineer | Python (FastAPI/Django) · Distributed Systems · PostgreSQL · AWS · Docker | Scalable APIs & Async Workflows | LLM Applications (RAG, Agent workflows, n8n)

About

I’m a Software Engineer focused on building reliable backend systems and production-grade AI application workflows. At Eli Lilly, I worked on end-to-end system development—from API design and data modeling to deployment and production monitoring—while integrating LLM-driven workflows into enterprise systems. My core focus is backend reliability and distributed workflows: designing modular APIs, modeling relational data, building async services, and implementing concurrency-safe transactional workflows that real production systems depend on. Recently I’ve been building LLM application workflows and orchestration systems, including RAG pipelines, embeddings, document chunking, vector search, and tool-based agent workflows (LangGraph), integrating them into production systems with emphasis on evaluation, reliability, and predictable API behavior rather than prompt-only experimentation. What I work on: • Backend Engineering — Python (FastAPI / Django), PostgreSQL, async services, distributed workflows, modular API design • Full-Stack Development — React, Next.js, Vue, TypeScript, Zustand, React Query • LLM Application Engineering — RAG pipelines, embeddings, vector search, agent workflows (LangGraph) • Cloud & DevOps — AWS (Lambda, API Gateway, S3), Docker, ECS Fargate, CI/CD with GitHub Actions Open to backend engineering, backend-leaning full-stack, and applied AI / LLM engineering roles in product-focused and remote-friendly teams. Portfolio: https://apoorv-x12.github.io/portfolio/

Location

India

AI stack

hugging face

langchain

langgraph

mcp

n8n

Synthesia

Work experience

AI Engineer at Outlier, Apr 2026 - Jul 2026 · 4 mos, Remote, Selected for Outlier’s Managed Workforce (MWF) Program and the INQA India program, contributing to large language model (LLM) evaluation and alignment through structured human feedback. Key contributions: - Evaluated and ranked LLM-generated responses for correctness, factual accuracy, reasoning quality, relevance, safety, and instruction adherence. - Designed and refined prompts to test model behavior across diverse scenarios, edge cases, and failure modes. - Identified hallucinations, logical inconsistencies, unsupported claims, instruction-following failures, and policy/safety issues. - Provided structured feedback and comparative judgments to support the improvement of generative AI systems. - Applied analytical reasoning and domain-specific evaluation criteria to assess the quality of model outputs. - Contributed high-quality human evaluation data used in LLM improvement and alignment workflows. Skills: LLM Evaluation · Generative AI · RLHF · AI Alignment · Prompt Engineering · Prompt Design · AI Quality Assurance · Human Feedback · Analytical Reasoning

Health and well-being at Career Break, Feb 2026 - Mar 2026 · 2 mos

Software Engineer at Eli Lilly and Company, Jul 2023 - Jan 2026 · 2 yrs 7 mos, Bengaluru, • Owned end-to-end distributed backend services using FastAPI, from API design and data modeling to deployment and monitoring, supporting internal products used by multiple teams. • Designed concurrency-safe event-driven workflows, explicitly reasoning about request isolation, retries, and failure handling in multi-step async systems. • Optimized PostgreSQL queries (joins, batching, indexing) for large datasets (millions of rows), reducing backend latency by ~30–40%. • Diagnosed and resolved production issues involving timeouts, partial failures, and inconsistent data using logs, metrics, and careful system reasoning. • Deployed and operated cloud-native services using Docker and AWS (ECS Fargate, Lambda [Python, Node.js], API Gateway, S3) with CI/CD via GitHub Actions and monitoring through CloudWatch. • Integrated LLM-backed workflows (RAG, embeddings, vector search, LangGraph) into backend systems with emphasis on predictable APIs, evaluation, and reliability, not prompt-only experimentation. • Built complex React/Vue interfaces (dashboards, multi-step forms, graph- and flow-based UIs), managing client-side and server-side state for REST, async, and AI-driven workflows, and reducing end-to-end load times by ~25–30% while keeping backend contracts and data integrity central. • Made deliberate design tradeoffs between sync and async processing to balance latency, throughput, and reliability under burst traffic.

Full Stack Engineer at SmartInternz, Apr 2022 - Jul 2022 · 4 mos, Remote, • Mastered Agile Full-Stack Web Development using Java and the Spring Boot framework, improving project delivery speed by 20%. • Implemented the MVC model and developed REST APIs, optimizing backend response times by 30%. • Designed and deployed a Nutrition App with a monolithic Spring Boot backend and a separate frontend, integrating MySQL for data persistence.

Augmented Reality Software Engineer at mirrAR, Feb 2022 - Apr 2022 · 3 mos, Remote

Education

Bachelor of Technology - BTech, Computer Science and Engineering, Vellore Institute of Technology

Public links

apoorv-x12.github.io: https://apoorv-x12.github.io/portfolio/