Nikhil Adyapak

Nikhil Adyapak

Machine Learning Engineer • MLOps Specialist • ML Systems • Data Science @ UW-Madison

I build

AI Intern @ Ciroos (SF Bay Area) | MS Data Science @ UW-Madison | Ex-Senior Software Engineer @ Bosch Global Software Technologies

Open to opportunities

Seeking 2027 Full-Time Roles · Available June 2027

Machine Learning Engineering · ML Systems · Software · MLOps · Data Science

Email Me Resume

Impact by the numbers

0% Faster Retraining Cycles
(Ray on AKS)
0M+ Vectors in
Image-Search Engine
0% Edge-Case Triage
Time Reduction
0+ Users on Private
On-Prem LLM
0+ ADAS Teams
Supported

01 About Me

Who I am

I am a Master's in Data Science student at UW-Madison, currently an AI Intern at Ciroos (SF Bay Area), where I run performance evaluations of AI-driven SRE agents across cloud environments, and built a Graph RAG context engine that grounds agent answers in incident history.

Before grad school I spent 2+ years at Bosch Global Software Technologies as a Senior Software Engineer on ADAS, part of the Automated Driving Alliance between Bosch and CARIAD, building MLOps platforms, distributed training on Ray + Azure Kubernetes Service, and scene-understanding data pipelines.

I like the engineering side of AI, not the demo. The part I care about is making it actually work in production.

Now building

Graph RAG @ Ciroos

Temporal knowledge-graph context engine that grounds SRE-agent answers in past incident history.

Based in

SF Bay AreaCiroos · Pleasanton, CA

Madison, WIUW-Madison

AI Agent Evaluation

Evals of AI SRE agents across cloud environments; RCA instruction pipelines; a Graph RAG context engine and a graph-free RAG alternative.

MLOps & Distributed Training

Ray Clusters on AKS cutting retraining from 4 weeks to 1 week; CI/CD with Terraform, Helm, Argo, GitHub Actions.

Large-Scale Retrieval

60M+ vector image-search backends (FastAPI + CLIP + ElasticSearch) curating fine-grained ADAS datasets.

2+Years Industry
2Publications
3.68GPA (MS)
1/10Qualcomm Track Teams

02 Professional Journey

All roles below, most recent first. Click any stop to jump to it.

AI Intern · Ciroos Current

Pleasanton, CA (SF Bay Area)

May 2026 - Dec 2026 8 mos
  • Productionized a Graph RAG context engine on a temporal knowledge-graph that sources SRE agent incident answers from team chat and investigation history.
  • Prototyped a graph-free RAG alternative that significantly lowered data ingestion cost over incident chats.
  • Conduct performance evaluations (evals) of AI-driven Site Reliability Engineering (SRE) agents by simulating faults across cloud environments.
  • Author and fine-tune detailed instruction pipelines that guide AI agents to accurately perform Root Cause Analysis (RCA) and fault diagnosis across varied system domains.
Graph RAGRAGAI AgentsEvalsSRERCA

Graduate Teaching Assistant · UW-Madison

CS564: Database Management Systems · Madison, WI

Jan 2026 - May 2026 5 mos
  • Led assignment sections and mentored 200+ students in database design, SQL, and Python; assisted instructors with automated grading pipelines and course logistics.
SQLPythonTeaching

Senior Software Engineer (AI Systems for ADAS) · Bosch

Bosch Global Software Technologies (BGSW) · Bengaluru, India · Automated Driving Alliance between Bosch and CARIAD

Aug 2023 - Aug 2025 2 yrs
MLOps Distributed Training Core ML Vector Search
  • Distributed ML Training: Designed and scaled a Ray Cluster system on Azure Kubernetes Service (AKS), reducing retraining cycles from 4 weeks to 1 week (75% faster) for 30+ teams.
  • MLOps Platforms: Automated ML pipeline CI/CD workflows on AKS (Terraform, Helm, Argo, GitHub Actions); integrated Prometheus alerting, accelerating experiment setup from 3 days to 5 minutes.
  • Large-Scale Retrieval: Architected a containerized FastAPI backend for a 60M+ vector image-search engine (CLIP embeddings + ElasticSearch), reducing edge-case triage by 95% (4 hours to 10 mins).
  • Core Machine Learning: Enhanced ADAS image dataset resolution by 4x utilizing ESRGAN transfer learning, and benchmarked Occlusion Estimation models on the nuScenes dataset via custom LiDAR-depth evaluation pipelines.
  • Scene Understanding: Engineered retrieval pipelines using foundation models (Mask2Former, Depth-Anything, OWL-ViT, CLIP) to generate scene graphs for complex ADAS corner cases.
  • Private LLM Access: Deployed on-prem Ollama LLM with a Python FastAPI wrapper, enabling team-wide LLM access (150+ users) with zero external data egress.
  • Cost Optimization: Implemented CLIP embedding reuse with ~80% cross-version alignment to avoid full re-indexing, and introduced image-complexity-based cost-aware labeling, cutting annotation costs by ~30%.
RayAKSTerraformMLflowCLIPElasticSearch

Engineering Internships · Bosch

MLOps Intern (Jan - May 2023) · Summer Intern (Jun - Jul 2022) · Bengaluru, India

2022 & 2023 2 stints · 7 mos
  • MLOps Intern: Built on-premise MLOps pipelines (DVC + MLflow) and accelerated multi-GPU object detection training (YOLOv5, Detectron2) on Azure ML Studio, cutting setup time from 1 week to 1 day (85% faster).
  • MLOps Intern: Configured Azure DevOps CI/CD to automate Docker deployments to Azure Container Registry; presented the POC at the internal Data Engineering Marketplace.
  • Summer Intern: Designed MLOps pipelines for object detection (YOLOv5, Detectron2) with DVC; implemented ensemble strategies and automated evaluation reports for autonomous driving robustness.
DVCMLflowAzure DevOpsDocker

Teaching Assistant · PES University

CS203: Statistics for Data Science · Bengaluru, India

Sep 2022 - Dec 2022 4 mos
  • Mentored 120+ students with hands-on sessions on Hypothesis Testing and Supervised ML; organized and judged departmental hackathons.
StatisticsPythonMentoring

Summer Research Intern · C-ISFCR, PES University

Center for Information Security, Forensics and Cyber Resilience · Bengaluru, India

May 2021 - Nov 2021 7 mos
  • Developed novel cryptanalysis methods for columnar transposition ciphers using Python & Optimization; published findings at IEEE SMARTGENCON 2022.
CryptanalysisPythonResearch

03 Featured Projects

Devpost Demo Repo

Sentinel-Edge AI

Selected for the Qualcomm Track at the UW-Madison MadData 26 Hackathon (1 of 10 teams selected from 350+ registrants). Air-gapped AI security auditor running 100% on-device (Snapdragon X Elite NPU). Detects legal risks and code vulnerabilities using local Llama 3.2 & RAG without cloud data egress.

Llama 3.2Snapdragon NPURAG + ChromaDBLangChain
DOI Repo Experiments Dataset

Code Runtime Complexity Prediction Springer 2023

Predicted Big-O runtime complexity of C/Java/Python code using static analysis and a BiLSTM over AST graph embeddings on the IBM CodeNet dataset, reaching 96% accuracy.

PythonTensorFlowScikit-LearnNetworkX
DVC Pipeline Detection Repos

MLOps POC for Pedestrian Detection

On-premise MLOps starter template with DVC + MLflow + Detectron2 covering the end-to-end ML lifecycle: versioned data, reproducible DAG pipelines, experiment tracking, and evaluation reports.

DVCMLflowPyTorchStreamlit
Private Work

Distributed ML Training System

Designed a Ray Cluster system on Azure Kubernetes Service, reducing retraining cycles by 75% (4 weeks to 1 week) for 30+ ADAS teams at Bosch.

RayKubernetesAzure
Private Work

Scene Understanding Pipeline

Engineered an image retrieval pipeline using foundation models (Mask2Former, Depth-Anything, OWL-ViT, CLIP) to generate scene graphs, curating fine-grained datasets for ADAS corner cases.

Hugging FacePyTorchTransformers
DOI Repo

Decrypting Transposition Ciphers IEEE 2022

Optimization techniques to break columnar transposition ciphers without prior knowledge of the encryption key, using matrix permutations and dictionary matching.

PythonCryptanalysisIEEE
Repo

Big Talk 1-hour speed build

Built in under 1 hour at the UW-Madison Claude Hacks, using Claude AI (Anthropic SDK) and FastAPI. A real-time matching application that analyzes user interests to find non-obvious thematic connections between people in a room to skip small talk.

Claude AIFastAPIPython

04 Technical Skills

Languages

Python
C
C++
Java
SQL
R

ML & Statistics

PyTorch
TensorFlow
Hugging Face
Scikit-Learn
OpenCV
Matplotlib

MLOps & Workflows

Ray
MLflow
DVC
Argo
Azure ML Studio
Prometheus
Grafana

Cloud & Infrastructure

Azure
AWS
GCP
Kubernetes (AKS)
Docker
Terraform
GitHub Actions
Linux

Databases & Tools

ElasticSearch
PostgreSQL
MySQL
MongoDB
FastAPI
Flask
Streamlit
GitHub

Certifications

Deep Learning SpecializationCoursera · 2023
AZ-900: Microsoft Azure FundamentalsUdemy · 2024
Kubernetes for BeginnersUdemy · 2024
Machine Learning in ProductionCoursera · 2023

05 Education

University of Wisconsin-Madison

MS in Data Science

Sep 2025 - May 2027 GPA 3.68 / 4.00
Machine Learning Foundation Models Database Design Statistical Models Statistical Inference

PES University, Bengaluru

B.Tech in Computer Science & Engineering

Aug 2019 - May 2023 GPA 3.78 / 4.00
Machine Intelligence & Data Science Algorithms Big Data Cloud Computing

06 Beyond Code

Honors, Competitions & Recognition

UW Madison · 2026

MadData 26: Qualcomm Track

Selected as 1 of 10 teams from 350+ registrants for the Qualcomm Track. Built Sentinel-Edge AI on the Snapdragon X Elite NPU.

UW Madison · 2025

3 Minute Thesis (3MT)

Competed in UW-Madison's 3 Minute Thesis, distilling autonomous-driving research for a non-technical audience.

Bosch · 2024

Top Achiever Award

Ranked 1st among 150+ engineers in Data Engineering at Bosch.

Bosch · 2024

Best Presenter Award

Recognized for technical excellence at the Bosch Data Engineering Department Meeting.

PES University · 2020-22

Distinction & MRD Scholarship

Distinction Award (Semesters 3-6) and the MRD Scholarship (40% tuition, top 20% of 1100+ students).

Leadership, Community & Volunteering

UW Madison · 2026

Wisconsin Experience Bus Tour (WEBT)

Selected for the fully-sponsored 5-day statewide leadership trip (35 students), engaging with industry, cultural, and community leaders across Wisconsin (UW-Madison SOLI).

Rhinelander, WI · 2026

Rhinelander Area Food Pantry

Volunteered during WEBT to organize, segregate, and stock food supplies, supporting local food security.

Bosch CSR · 2023-25

Tech Mentorship & Inclusion

Mentored candidates with disabilities in interview preparation (Youth4Jobs) and taught Computer Science fundamentals to underprivileged students.

Bosch & PES CSR

Animal Welfare & Sustainability

Supported habitat restoration for rescued elephants, dog adoption drives (CUPA), and campus afforestation planting saplings for green cover.

Bosch CSR · 2023-25

Social Impact

Assisted agricultural initiatives where produce revenue funded monthly stipends for differently-abled individuals.

Bosch & PES CSR

Community Outreach

Created interactive educational materials for rural childcare centers, coached sports, and helped renovate government-school facilities for underprivileged students.

07 Get in Touch

Seeking 2027 full-time roles, available June 2027. Also happy to talk evals, distributed training, and retrieval at scale.

Email

nikhiladyapak31@gmail.com

Best way to reach me. I usually respond within 12 hours.

Send Email

LinkedIn

in/nikhil-adyapak

GitHub

@NikhilAdyapak

Google Scholar

Publications
Email copied to clipboard