Seattle, WAOpen to AI roles

Rithesh Bejjarapu

Staff AI Engineer

Data→Inference→Agents→Production

Staff AI Engineer with 10+ years of experience across software engineering, data platforms, and production AI systems, including years embedded with client and business teams as a consultant. Specializes in local and open-source LLM inference, RAG, agentic systems, and enterprise knowledge platforms, translating ambiguous business problems into shipped AI that delivers measurable impact through decision intelligence, automation, and cost optimization.

Resume LinkedIn GitHub Contact

$2.1M: revenue at risk surfaced; AI data analyst, Slide
44%: faster security assessments; AI + human-in-the-loop, 9 to 5 days
50%: external AI spend cut; hybrid local inference
10+: years shipping for clients; AI, data, and delivery

A T-shaped engineer

Broad foundation across the top, from client delivery and stakeholder alignment to systems engineering. Deep AI-engineering specialization down the stem. Hover or tap any node to see how it fits.

Breadth

Depth / AI Engineering

Depth

Local Inference

Serving quantized open-source models on-prem, routed by task complexity, privacy, latency, and cost.

50%

external AI spend cut

Where

Slide

vLLMllama.cppQuantized Models

Experience

10+ years across production AI, data platforms, and enterprise software, with the most recent work centered on end-to-end AI systems.

Slide

Staff AI / Data Engineer

Nov 2025 to PresentRemote

BigQueryCompany BrainvLLMllama.cppMCPModel Routing

Data Foundation & Decision Intelligence

Architected the company's governed data foundation on BigQuery, transforming fragmented operational, financial, and partner data into reusable models and data products powering executive analytics, AI applications, and agentic workflows.
Built an evaluated AI data analyst that autonomously investigates financial anomalies, trends, and partner signals; surfaced $2.1M in revenue at risk and generated opportunity signals associated with 75% of successful upsell outcomes during the evaluated period.

Company Brain & Agentic Platform

Designed a shared company brain that unifies structured and unstructured organizational knowledge into a common context layer for employees, AI assistants, and agents, with evaluation and recursive improvement loops that use failures and feedback to improve retrieval and response quality.
Built an internal agentic platform connecting the company brain, data warehouse, and operational systems through MCP-based tools, enabling context-aware automation with RBAC-controlled access to sensitive data and actions.

Local Inference & Model Systems

Designed a hybrid inference architecture spanning specialized SLMs, locally hosted open-source LLMs, and frontier cloud models, routing workloads based on task complexity, quality, privacy, latency, and cost.
Deployed and operated 6+ open-source models with vLLM and llama.cpp, routing 60%+ of eligible internal workloads to local inference and reducing monthly external AI spend by 50%, while preserving frontier models for workloads requiring higher reasoning capability.

Amazon

Sr. SDE, AI & Security Automation

Sep 2022 to Oct 2025New York, NY

LLM SystemsOSINTRisk IntelligenceAWSOrchestration

AI-Assisted Security Assessment

Architected an LLM-based security assessment system that evaluates third-party engagements against assessor runbooks, Amazon security policies, vulnerability intelligence, and internal security data, generating grounded initial assessments for human reviewers and reducing average assessment closure time from 9 days to 5 days within four months.
Built an AI-powered customer intelligence system that aggregates OSINT, cybersecurity, fraud, vulnerability, and internal risk signals into structured risk profiles; evaluated 300+ companies and surfaced 45 previously unidentified potential risk signals for assessor investigation.

Assessment Platform & Automation

Designed and built an AWS-based self-service assessment platform that dynamically assembles questionnaires, evidence requests, and evaluation workflows based on engagement context, replacing static assessment flows and improving the consistency and accuracy of security posture analysis.
Designed middleware and orchestration services on AWS to connect the assessment platform with internal security systems, normalizing data and coordinating asynchronous workflows across multiple APIs and downstream services.

JPMorgan Chase & Co.

Senior Software Engineer

Nov 2021 to Sep 2022Wilmington, DE

Low-Code PlatformWorkflow AutomationEnterprise APIs

Automation Platform Engineering

Built a governed low-code automation platform that enabled business teams to design, approve, deploy, and maintain workflows while securely accessing enterprise credentials, APIs, and internal systems through a controlled engineering layer.
Expanded the platform to support 120+ reusable actions and 50+ production workflows, enabling self-service automation across business teams and saving 300+ engineering hours previously spent on custom implementation and support.

Earlier engineering experience

AMC Networks

2020 to 2021

Senior Software Developer

Built IT support and employee lifecycle automations across onboarding, offboarding, and service operations, reducing manual effort for the Service Desk by approximately 15 hours per month.

Yale University

2019 to 2020

Application Developer

Developed IT asset management workflows supporting healthcare and enterprise operations, and built a mobile experience enabling service desk agents to receive alerts and take operational actions from the field.

7-Eleven

2018 to 2019

Software Developer Consultant

Built workflow and support automations for enterprise IT operations, improving request handling and reducing manual coordination across distributed teams.

USAA

2016 to 2018

Software Developer Consultant

Worked within a core engineering team building and maintaining enterprise automation for access, provisioning, and internal engineering workflows used across corporate technology teams.

Founder & creative AI

Applying the same practical AI systems thinking outside of enterprise engineering.

Svaram.ai

Founder, Side Venture · Oct 2025 to Present

An AI-assisted music studio for custom business anthems and promotional music, combining generative models with artist-in-the-loop creative direction to produce commercially usable tracks.

AI-enabled production workflows

Music generationStem separationVocal isolationKaraoke creationPost-production automation

Regionly

getregionly.com

AI-powered competitive intelligence product that analyzes public business data and generates localized market insights and actionable recommendations.

Technical skills

AI Engineering

Local & Open-Source LLMsRAGAgentic SystemsModel RoutingLLM EvaluationContext EngineeringEmbeddingsVector SearchGuardrailsAI ObservabilityMCP

Inference & Model Systems

vLLMllama.cppSmall Language ModelsModel ServingQuantized ModelsLocal InferenceHybrid Local & CloudLatency & Cost Optimization

Data & AI Platforms

BigQueryData PipelinesData ModelingEnterprise Knowledge SystemsStructured & Unstructured DataRetrieval InfrastructureAnalytics Data ProductsOpenSearch

Software & Cloud

PythonJavaJavaScriptREST APIsAWS BedrockSageMakerLambdaAPI GatewaySQLNoSQLDockerCI/CDDistributed Systems

Education

M.S. in Computer Science

Northwest Missouri State University · Missouri

2014 to 2015

B.E. in Computer Science

Vasavi College of Engineering · Hyderabad, India

2010 to 2014

Writing

Notes from the build: local inference, retrieval, agents, evals, and the data work underneath.

Jul 14, 20263 min read

Notes on building AI systems that survive production

Why I started writing here, and the kinds of things I want to document: local inference, agents, evals, and the unglamorous data work underneath it all.

View all writing

06 / Contact

Let's build AI systems that work in production.

Open to AI roles, from building end-to-end AI platforms to forward-deployed delivery embedded with your team. Reach out and let's talk.

rithesh.bejjarapu@gmail.com

LinkedIn GitHub Resume