👋 Hi, I'm Smruti Singh
Customer-facing engineer with experience driving end-to-end implementations, integrations, and post-deployment optimization. Strong at translating ambiguous workflows into scalable technical solutions with a product mindset. I bring an ownership-driven approach across building, debugging, and stakeholder communication to ensure adoption and impact.

About Me
I approach engineering the same way I approach Bharatanatyam through structure, rhythm, and precision. Years of classical dance taught me discipline, pattern recognition, and how small misalignments can break an entire system. That mindset now shapes how I approach customer environments: beneath the complexity, there’s always a workflow that can be understood, improved, and rebuilt.
I’ve worked in customer-facing engineering roles where I’ve led end-to-end implementations, integrating enterprise systems (SSO, LDAP, APIs) and taking deployments from kickoff to production. My work focuses on understanding real-world workflows, translating ambiguous requirements into reliable technical solutions, and ensuring systems don’t just go live but actually get adopted and deliver value.
I operate at the intersection of engineering and execution, debugging issues in real time, collaborating with stakeholders, and continuously optimizing systems post-deployment. I care about building solutions that are not just functional, but scalable, repeatable, and aligned with how teams actually work.
Outside of work, reading keeps me grounded. It reinforces that precision and creativity go hand in hand something I carry into every implementation I own: thoughtful, structured, and built with the end user in mind.
Education
Master of Science in Information Management
University of Illinois Urbana-Champaign
Aug 2024 - May 2026
GPA: 3.93/4.0
Bachelor of Technology in Information Technology
University of Mumbai
Aug 2018 - May 2024
GPA: 9.16/10
Work Experience
Software Engineer
University of Illinois Foundation
May 2025- May 2026
- Owned an end-to-end data pipeline from raw CRM data to decision-ready segments - replaced 30+ ad-hoc SQL queries with a single Python ETL engine on a serverless AWS stack (API Gateway → SQS → Lambda), cutting donor list turnaround from 24 hours to 6
- Built and operated a production integration with a third-party enterprise system (Blackbaud CRM → AWS API Gateway) - added retry logic and structured error handling so campaign uploads of 10,000+ records processed reliably without manual babysitting
- Designed data-quality safeguards directly into the pipeline - SQL validation views across 60+ campaign segments to catch duplicate donor records before they reached fundraising teams, cutting duplicates from ~5% to ~1%
- Shipped changes safely at production scale - built automated test suites and GitHub Actions CI/CD pipelines enabling zero-downtime deploys, reinforced by peer code review that cut post-deploy defects by 40%
- Worked directly with fundraising stakeholders to translate campaign requirements into pipeline logic, ensuring segmentation criteria matched how the team actually made outreach decisions
Software Engineer - Implementation
ARCON TECH Solutions
June 2022 - June 2024
- Served as the forward-deployed technical owner for 15+ enterprise PAM deployments - led architecture reviews mapping each client's network, firewall, and server environment before onboarding, enabling 200+ devices per engagement without production surprises
- Integrated PAM directly into each client's existing identity stack - connected Active Directory and AWS Secrets Manager to extend secure access to 250+ privileged users per client
- Debugged and hardened SSO-based authentication in live client environments - mapped SAML assertions and LDAP structures into working SSO configurations, catching identity-provider misalignments before they caused downstream access failures
- Identified a recurring operational pain point across multiple clients and built the fix unprompted - a Lambda-driven log archival pipeline on S3 that cut forensic log lookup time from 20 minutes to 5 and reduced 90-day storage costs by ~30%; presented the design and cost case directly to client leadership for sign-off
- Automated deployment consistency to reduce manual configuration risk - built reusable Python tooling with monitoring to catch config drift, cutting deployment time from 5 days to 3
- Acted as the technical bridge between customers and the product team - partnered with sales pre-close to scope proof-of-concepts around real client pain points, then owned delivery and stabilization of those same engagements post-close
- Stayed embedded with clients post-launch through a scheduled escalation rotation, iterating until each environment reached steady-state operation independently
Skills
Languages
Cloud
Tools & Frameworks
Databases
AI & ML Systems
Identity Management
Soft Skills
Certifications
My credentials from industry-leading organizations.
AWS Certified Solutions Architect - Associate
Amazon Web Services - 2025
AWS Certified Cloud Practioner
Amazon Web Services - 2025
HackerRank SQL (Intermediate) Certificate
HackerRank - 2025
Projects
MedScribe RAG
Designed a two-stage LLM pipeline turning raw EMS call audio into validated insurance billing records - speech transcription (Whisper) → structured medical data extraction → automated insurer matching - reaching 96% field accuracy across 9 real-world call scenarios
EdgeMind - Offline-First On-Device RAG AI Assistant
Built an offline-first Android AI assistant that runs Llama 3.2 3B entirely on-device using NexaSDK with Qualcomm Snapdragon NPU acceleration and automatic CPU/GPU fallback for non-Elite devices.
OrderPilot
Architected a kiosk and mobile application using React and TypeScript, enabling touch and voice-based interactions.
SnapSplit - AI Powered Receipt Parsing & Expense Splitting
Built a full-stack AI-powered receipt parsing and expense splitting application that uses GPT-4o Vision for structured extraction from receipt photos, with confidence scoring and a human-in-the-loop review and correction workflow.
Netflix Churn Analysis
Defined churn from user inactivity and used exploratory data analysis to surface the strongest risk signals - engagement level and favorite-genre patterns.
Credit Card Fraud Detection
Framed fraud detection around behavioral signals rather than transaction size - engineered features for transaction velocity, device/IP changes, and timing patterns (night-time, weekend, prior-fraud history) to capture sequential intent, inspired by behavior-trace classification research from JPMorgan Chase AI.
Leadership & Initiatives
Beyond coding, I am driven by a desire to lead initiatives that create meaningful social impact and empower the communities around me.
Course Assistant, GIES-Business 401
Led and facilitated 3 weekly small-group sessions (10–20 students each), creating an inclusive environment for collaboration and learning.
Research Assistant, CHIME
Quantified the correlation between fluctuating climate variables and regional mortality rates to provide data-driven insights for public health policy.
Curriculum Head, ICHCHHA
Worked as a subject curriculum head where designing question papers, outline of portion for standard 8th and 9th.
Get in touch
Have a question or want to work together? Feel free to reach out.
San Francisco, CALIFORNIA
I am available for full-time opportunities.