AI/ML ENGINEER · PRODUCTION GENAI

I build AI systems that survive production.

I’m Abhinav Reddy Padamati, an AI/ML Engineer focused on multi-agent applications, RAG, evaluation, distributed inference, and cloud-native ML systems. I care about the part after the demo: quality, latency, cost, security, and reliability.

Explore my work ↗Contact me
Databricks Mosaic AILangGraphMLflowAWS / Kubernetes
groundedness ↑ 26%
gpu cost ↓ 31%
agent_runtime.traceLIVE
01
Intent Routerclassify + route user request
ready
02
Hybrid Retrievalvector + keyword + metadata
ready
03
Cross-Encoder Rerankprecision-first candidate scoring
ready
04
Tool / MCP Agentgoverned actions + validation
ready
05
Evaluate + ServeMLflow evals + observability
ready
25K+user environment
37%latency reduction
29%unsupported ↓
5+Years across ML & GenAI
25K+Enterprise user environment
26%Grounded-response accuracy gain
31%GPU inference cost reduction
01 · EXPERIENCE

Built for the messy
part of AI.

Production AI is not one model. It is retrieval, orchestration, serving, evaluation, observability, infrastructure, and the trade-offs between them.

JAN 2025 — PRESENT

AI/ML Engineer

Databricks · San Francisco, CA
Built multi-agent enterprise AI workflows supporting a 25,000+ user environment and improving grounded-response accuracy by 26%.
Engineered hybrid RAG with metadata filtering, semantic chunking and cross-encoder reranking, reducing unsupported responses by 29%.
Optimized FastAPI + Model Serving + vLLM + Ray inference paths, reducing end-to-end latency by 37% and GPU inference cost by 31%.
Created MLflow evaluation and OpenTelemetry observability for groundedness, safety, tool accuracy, regression detection and production diagnosis.
Databricks Mosaic AILangGraphMLflowVector SearchFastAPIvLLMRayKubernetesOpenTelemetry
MAR 2021 — JUL 2024

Machine Learning Engineer

Accenture · India
Built cloud-capacity forecasting models with XGBoost, Prophet and Scikit-learn, improving forecast accuracy by 23%.
Developed PySpark feature pipelines that contributed to a 27% reduction in overprovisioned cloud infrastructure costs.
Operationalized forecasting through async FastAPI services on AWS, reducing inference latency by 35%.
Built Kafka + Spark streaming, drift monitoring, retraining, CI/CD and GitOps workflows for reliable production ML operations.
XGBoostProphetPySparkKafkaAirflowTerraformArgoCDEvidently AI
02 · FEATURED WORK

Systems, not demos.

Projects selected to show architecture, reliability, measurable impact, and end-to-end ownership.

FEATURED / 01

Enterprise AI Knowledge Assistant

Enterprise RAG application using FastAPI, LangChain, FAISS and Sentence Transformers for semantic search and document Q&A across PDF, DOCX and TXT sources.

92% answer relevance+30% precision500+ concurrent queries<2s avg response
FastAPILangChainFAISSSentence TransformersPostgreSQLAWS S3
FEATURED / 02

AWS Event-Driven Notification System

Fault-tolerant serverless notification platform built around asynchronous event processing, retries, queues, dead-letter handling and monitoring.

15K+ events99.8% delivery success90% failed deliveries ↓60% detection time ↓
LambdaEventBridgeSQSSNSDynamoDBAPI Gateway
03 · TOOLKIT

My production stack.

Focused on the layers needed to build, ship, observe, and improve modern AI products.

GenAI / Agents

RAG, LangChain, LangGraph, Vector Search, Hybrid Retrieval, Reranking, Metadata Filtering, LoRA / QLoRA, DBRX, OpenAI, Claude, Gemini

ML / Deep Learning

PyTorch, TensorFlow, Scikit-learn, XGBoost, LightGBM, Prophet, TFT, model evaluation

MLOps / Platform

MLflow 3, Mosaic AI, Model Serving, Unity Catalog, Databricks Workflows, Evidently AI, monitoring

Serving / Distributed

FastAPI, Ray, vLLM, Redis, asynchronous processing, microservices, MCP servers, distributed GPU inference

Data Engineering

Spark, PySpark, Delta Lake, Kafka, Structured Streaming, Auto Loader, Airflow, Databricks SQL

Cloud / DevOps

AWS, Kubernetes, Docker, Helm, Terraform, GitHub Actions, Jenkins, ArgoCD, GitOps

Observability / Security

OpenTelemetry, Prometheus, Grafana, Alertmanager, OAuth2, RBAC, Trivy, SonarQube

Languages

Python, SQL, Bash, PySpark plus practical API, systems and automation development

04 · CREDENTIALS

Education & certifications.

Education

Master of Science in Computer ScienceUniversity at Buffalo
Bachelor of Science in Computer SciencePadmabhooshan Vasantdada Patil Institute of Technology (PVPIT)

Certifications

Databricks Certified Generative AI Engineer Associate
Certified Kubernetes Administrator (CKA)
AWS Certified Machine Learning Engineer – Associate
OPEN TO AI/ML & SOFTWARE ENGINEERING ROLES

Need someone who can take AI from prototype to production?

I’m interested in roles where GenAI, ML systems, distributed serving, evaluation, and cloud infrastructure come together in one real product.

padamatiabhinavreddy@gmail.com ↗(716) 547-9802LinkedIn ↗Download Resume ↓