Looking for a job

Abhishek Srivastava

@srivabhi22

Message

I build production LLM and multi-agent systems at scale.

India

Message

What I'm looking for

I'm looking to build production AI systems where I can own retrieval, agent design, evaluation, deployment, and monitoring end to end, with opportunities to solve high-scale, measurable product problems.

I've built production LLM, RAG, and multi-agent systems for Arivihan Technologies, serving 120K+ students monthly. I replaced manual doubt resolution with a six-agent tutoring system that reduced human escalations from 18% to 6%.

I design retrieval, evaluation, deployment, and monitoring end to end. At Arivihan, I built memory and serving systems with FastAPI, Redis, DynamoDB, and SQS that sustained 1.2K requests per second at p99 850ms.

Previously at InfraDock AI and UnifyApps, I built document-ingestion, Graph RAG, and enterprise search systems across PDFs, Excel files, images, and 20+ connectors. My work improved retrieval accuracy, reduced query latency, and made over 100K enterprise documents searchable.

I'm also interested in reliable agentic workflows, having built a LangGraph question generator with a 96% valid-plan rate and deployed vision-based terrain navigation for a Unitree Go2 at IIT Kanpur.

Experience

Work history, roles, and key accomplishments

Junior Machine Learning Engineer

Arivihan Technologies Pvt. Ltd.

Mar 2026 - Sep 2026 (6 months)

• Developed a 6-agent hybrid RAG system (30+ tools) for 120K+ monthly users, cutting escalations by 35%.
• Built a FastAPI, Redis, and SQS backend handling 1.2K req/s, reducing session loss from 4.5% to 0.3% via DLQ.
• Built ClickHouse based dashboards across 20M+ messages and automated 800+ call audits/day saving 25
hrs/week in QA.

Python PyTorch LangGraph fastAPI AWS Redis DynamoDB RAG Multi Agent Systems LLM OpenAI Gemini Clickhouse

AI Systems Engineer Intern

InfraDock AI

Sep 2025 - Feb 2026 (5 months)

• Developed Graph RAG pipelines on Qdrant resolving enterprise queries, boosting retrieval accuracy by 35%.
• Built LLM ingestion pipelines in PostgreSQL to extract structured data from PDFs/Excels, hitting 92% accuracy
across 5K+ documents.

Agentic Workflows JSON RESTful API Design Python Vector Databases Data Pipeline Architecture Data Ingestion LLM APIs Graph RAG

Product Engineer Intern

UnifyApps

May 2025 - Jul 2025 (2 months)

• Engineered enterprise RAG services spanning 20+ connectors to make 100K+ documents searchable.
• Developed RBAC-aware agentic indexers with semantic pre-filters that reduced p95 latency by 25% to <200ms.
• Shipped 150+ API integrations and benchmarked OCR agents, driving a 15% lift in extraction accuracy.

Groovy Agentic Workflows RBAC JSON RESTful API Design

Education

Degrees, certifications, and relevant coursework

Indian Institute of Technology Kanpur

Bachelor of Technology, Mechanical Engineering

2022 -

Activities and societies: Minor in Machine Learning & Applications; Coursework: Machine Learning, Statistical Natural Language Processing, Data Structures & Algorithms, Probability & Statistics

Pursuing a B.Tech in Mechanical Engineering with a minor in Machine Learning and Applications. Achieved AIR 2889 in JEE Advanced 2022 among 1.02M candidates and AIR 122 in NEST 2022.