👋 Hello, I’m
Rohan Patil
Building AI Systems
That Scale 🚀
AI/ML Engineer with experience at Perplexity and Amazon, building production-grade LLM pipelines, RAG systems, and distributed ML infrastructure for real-world high-scale environments.
5+ yrs
Experience
25%
Latency Improvement
1M+
Requests handled
Selected Projects
Adaptive RAG Chatbot
Dynamic retrieval + query routing system for grounded LLM responses.
↓ Latency 40% | ↑ Accuracy 25%
Resume → Job Matching Agent
Agentic AI system for matching resumes to job descriptions using embeddings.
↑ Match Precision 30%
Real-Time ML Pipeline
Kafka + Spark streaming pipeline for low-latency feature engineering.
1M+ events/day processed
LLM Evaluation Dashboard
Tracking model performance, latency, and hallucination metrics.
Improved evaluation accuracy
Vector Search Engine
Hybrid FAISS + Redis retrieval system for semantic search.
Recall ↑ 20%
Inference Optimization System
Optimized GPU inference using batching and Triton.
Throughput ↑ 25%
Experience
AI/ML Engineer — Perplexity
June 2024 – Present · San Francisco, CA
- • Architected RAG pipelines integrating vector search + web indexing.
- • Built FAISS + Redis hybrid retrieval improving recall/precision tradeoff.
- • Optimized Triton GPU inference → +25% throughput.
- • Designed LLM routing (on-device + cloud) for sub-second latency.
- • Improved factual consistency via ranking + citation pipelines.
- • Built evaluation systems tracking latency, accuracy, UX metrics.
- • Led 0→1 agentic AI features → +18% engagement.
AI/ML Engineer — Amazon
Oct 2019 – June 2023 · India
- • Built batch + streaming pipelines using AWS, Spark, Kafka.
- • Designed feature systems → +30% faster data access.
- • Prevented training-serving skew in real-time ML systems.
- • Built Kafka + Spark streaming pipelines for low latency updates.
- • Orchestrated ML workflows with Airflow + SageMaker.
- • Built drift detection + monitoring datasets.
- • Reduced infra cost by ~15% via optimization.