Open to opportunities

abhii raj

@abhiiraj

Message

I build cloud-native AI platforms with LLM governance, observability, reliability, and cost control.

India

Message

What I'm looking for

I'm looking to build safe, reliable generative AI platforms in an embedded, ego-free cross-functional team. I'm available immediately and open to relocating to the UK with visa sponsorship.

I'm building AI platforms at ZS Associates, including an async document processing system that handles 12,000+ pages per session at 8x lower cost than OCR APIs. I also architected a Graph RAG pipeline that reached 91% extraction accuracy on complex layouts.

Previously, as Founding Lead AI Engineer at Neural Vault, I deployed full-stack AI observability on AWS EKS at 99.9% uptime and built agents and copilots supporting three government contracts across 1M+ citizen records. My governance and evaluation work reduced hallucinations by 40%, while an AI voice bot automated 60% of Tier-1 support.

At Arcesium, I built Python and Java services for $50B+ AUM FinTech portfolios, processing 5M+ daily records with SOC2-compliant auditability. I improved MTTR by 45%, cut p99 API response time by 70%, and led architecture, code quality, and observability standards for a 40-developer team.

I bring 8+ years across cloud infrastructure, backend systems, AI agents, and production LLM operations, with hands-on experience in AWS, Terraform, Kubernetes, Datadog, LangSmith, and AgentOps.

Experience

Work history, roles, and key accomplishments

Current

Engineering AI Lead

Current

ZS Associates

Apr 2026 - Present (4 months)

Built an async document processing platform handling 12,000+ pages per session at 8x lower cost than OCR APIs, with cost controls that cut regulatory prep by four weeks per cycle. Architected a Graph RAG pipeline reaching 91% extraction accuracy, outperforming AWS Textract and Azure Document Intelligence by 2.8x on complex layouts.

Python AWS Neo4j Redis Celery Graph RAG

Founding Lead AI Engineer

Neural Vault

Apr 2025 - Mar 2026 (11 months)

Implemented full-stack AI observability with AgentOps and LangSmith for tracing, reliability, and cost efficiency, deployed on AWS EKS at 99.9% uptime. Built AI agents and copilots with LangChain, LangGraph, and MCP servers using tool-use architectures as the sole AI engineer, powering 3 government contracts across 1M+ citizen records.

Python Docker Kubernetes Langsmith AgentOps MCP Lang Chain LangGraph fastAPI Redis

Senior Software Engineer, Python & Java Backend

Arcesium

Feb 2019 - Mar 2025 (6 years 1 month)

Designed Datadog-integrated observability with distributed tracing and streamlined CI/CD via Terraform and GitHub Actions, reducing MTTR by 45%. Built a Redis caching and rate-limiting layer that cut p99 API response time by 70%, and reduced critical query latency by 99% (2 min to 200ms).

Python Java AWS Terraform Docker Datadog GitHub Actions Kafka Redis PostgreSQL fastAPI Spring Boot

Software Developer, Python / ML

SAP Labs

Jan 2018 - Jan 2019 (1 year)

Built an ML model with an end-to-end ETL pipeline for inventory management, achieving a 30% reduction in spoilage cost, and optimised SAP HANA SQL to cut query time 40%.

Python scikit learn PyTorch Pandas NumPy SAP Hana ETL

Education

Degrees, certifications, and relevant coursework

Indian Institute of Information Technology, Allahabad

Bachelor of Technology, Information Technology

2014 - 2018

Grade: 8/10 CGPA

Activities and societies: IIT JEE Rank 4698 (Top 1%), 3x Science Olympiad School Topper

Pursued a Bachelor of Technology in Information Technology, achieving a CGPA of 8/10. Ranked 4698 in IIT JEE (top 1%) and was a 3x Science Olympiad school topper.