himnish

Work & experience

Work & {experience}

A decade of curiosity, condensed: production AI at Harness, internships across fintech, machine learning, and consulting, and years of leadership at UIUC. Each role below opens to the full detail. Start with the previews, dig in where you like.

Download full résumé (PDF)
Sept 2025 – present

Building AI-driven test automation from the ground up, and owning the production issues that block releases. Promoted to SE II (AI) after about eight months.

  • Built and shipped a Prompt Scoring & Enhancement System for AI-driven test automation: execution handlers, prompt-construction logic, and observability (logging + metrics) that made debugging and validation far easier across environments.
  • Owned end-to-end resolution of critical production and CI-blocking issues (intermittent 504 errors, AI logic inconsistencies): reproduced failures in staging, found root causes, and shipped fixes that stopped the repeat escalations.
  • Improved test reliability and observability by rethinking how test outcomes and warnings surface to engineers.
  • Streamlined developer workflows with deep-linking between test-suite runs and pipeline executions, cutting investigation time.
  • Migrated pipeline and test configurations to version-controlled infrastructure, reducing manual dependencies.
  • Accelerated release-pipeline performance through iterative experimentation and optimization.
  • Helped hold together infrastructure and monitoring for a brand-new AI Test Automation system under rapid development.
  • Partnered with customers during Proof of Value engagements, building what they needed and fixing issues fast to improve conversion.
Dec 2024 – Jun 2025

Automated financial reconciliation with SQL and Python, cutting manual effort and sharpening accuracy.

  • Automated disbursement-reconciliation workflows using SQL and Python, significantly reducing manual effort and improving the accuracy of financial operations.
  • Analyzed transaction-level financial data to spot discrepancies and quantify operational risk; ran exploratory data analysis on large-scale payment datasets to detect anomalies.
  • Documented reconciliation workflows and data-validation procedures; worked cross-functionally with Finance, Accounting, and Engineering.
Jun 2024 – Aug 2024

Shipped ML pipelines that detect drift in satellite imagery at 94% accuracy.

  • Deployed ML pipelines to detect drift in satellite imagery data with 94% accuracy.
  • Set up human-in-the-loop methods for ground-truth accumulation and automated grading, using active-learning techniques.
  • Automated data validation and monitored clusters with Grafana and Groundcover.
May 2023 – Aug 2023

Built a Next Best Action predictor with a Deep Q-Network: 93% precision on the next two actions.

  • Developed a Next Best Action predictor for individualized client preferences using a Deep Q-Network that predicted the next two actions with 93% precision.
  • Converted an internal NLP tool from SpaCy to FlairNLP for more optimized Named Entity Recognition.
  • Devised an algorithm to extract images from PDFs and assess their usability for billing and HR purposes.
May 2022 – Aug 2022

Built omnichannel retail workflows for MNC mall corporations, driving up to 90% higher customer retention.

  • Facilitated omnichannel retail strategies for MNC mall corporations, driving up to 90% higher customer-retention rates.
  • Built optimized download workflows for multiple file formats with a 25-person team.
Dec 2019 – Jan 2020

Analyzed HR data with MySQL and external APIs to surface the department's cost-to-productivity ratio.

  • Analyzed HR-department data trends using MySQL and external APIs; demonstrated the department's cost-to-productivity ratio and ways to improve it.

UIUC leadership & teaching · 2021 – 2025

Jan 2023 – present

Real-time transcription research for accessibility, under Professor Lawrence Angrave: a Whisper-based backend at ~2 seconds per sentence.

  • Worked under Professor Lawrence Angrave on ScribeAR, an open-source real-time captioning tool for deaf and hard-of-hearing students.
  • Implemented an alternate speech-to-text backend using OpenAI's Whisper (Whisper.cpp), cutting transcription time to ~2 seconds per sentence.
  • Built a React hook for Whisper API calls with options for softness levels, transcription speeds, and history management.
  • Compiled 100+ domain word lists to improve technical-word recognition in STEM classroom settings.
  • Researched speaker identification, diarization, and audio classification.
  • (See the full story on Research and Projects.)
Aug 2023 – May 2025

Lead ELA for CS in ENG 100, guiding 500+ first-year computer science students through their first semester.

  • Lead Engineering Learning Assistant for computer science in Engineering 100, the first-year orientation course, supporting 500+ first-year CS students.
  • Ran interactive sessions twice a week and gave personalized mentorship through students' first semester.
  • Created lesson guides and assignments, then graded them with individual, personalized feedback.
  • Cultivated a collaborative learning environment for students adjusting to university-level computer science.
Aug 2022 – May 2025

Supported 50+ residents; organized events with a team of 10 for a 550+ person hall.

  • Served as RA for 50+ students on a floor; held regular one-on-ones offering emotional and academic support; organized events with a team of 10 for a residence hall of 550+ people.
2021 – 2025

Illini Blockchain developer and ACM SIG-AIDA, a semester-long AI bootcamp.

  • Illini Blockchain: Developer (Mar 2022 – May 2025)
  • ACM SIG-AIDA (AI & Data Analytics) (Aug 2021 – May 2025), ran a semester-long AI bootcamp; trained ML models with PyTorch and the BERTweet tokenizer.