About
Early-career ML professional with experience across the data/ML lifecycle: problem scoping, data wrangling (Python/SQL), exploratory analysis, feature engineering, and modeling with scikit-learn, TensorFlow, and PyTorch. Comfortable designing metrics and experiments, turning insights into dashboards or APIs, and collaborating with engineers and stakeholders to ship measurable improvements. Looking to grow in roles where I can build reliable data pipelines, optimize models for latency/throughput, and communicate results clearly.
Open to:
Machine Learning EngineerAI/ML EngineerData Scientist IData Analyst I
Technical Skills
ML Frameworks
TensorFlowPyTorchscikit-learnXGBoost
Languages
PythonSQLRJavaHTMLCSSJavaScript
MLOps & Deployment
Model ServingQATDockerAPIsStreaming Inference
Tools
GitJupyterMongoDBFirebaseAngularAndroid StudioVS Code
Core Concepts
Feature EngineeringOptimizationInference PipelinesComputer VisionReal-Time Systems
Experience
AI Software Engineer
Splashgain Technologies
Jul 2023 – Mar 2024 • Pune, India
Stack: TensorFlow, Python, Angular, HTML/CSS/JS
- Hardened a production CBT proctoring pipeline (post-processing, alert routing, reviewer UX).
- Optimized TensorFlow serving (input pipeline tuning, batching, graph tweaks) for ~40% lower prediction latency and ~15% faster end-to-end processing across 5k+ active users.
- Integrated flags into the Eklavvya admin portal to improve reviewer throughput.
Software Intern
Splashgain Technologies
Jul 2022 – Jun 2023 • Pune, India
Stack: TensorFlow, EfficientDet, Python
- Developed and deployed EfficientDet-D0 object detection for CBT cheating cues (~92% test accuracy; processed 50k+ frames/day).
- Built streaming inference with quantization-aware training and real-time alert generation; automated flagging reduced manual invigilation by ~20 hrs/week.
- Added telemetry and error handling for stable rollout and faster triage.
Data Analytics Intern
Mahindra & Mahindra
Aug 2021 – Sep 2021 • Nashik, India
Stack: Python, MongoDB
- Analyzed 10M+ CNC logs; segmentation + anomaly scoring ≈95% accuracy.
- MongoDB aggregations cut detection time by ~35% for faster maintenance response.
Projects
Self-Refinement Loop for LLMs
Aug 2025
LLMsEvaluationAutomation
- Iterative self-critique with multi-critic scoring (clarity, completeness, alignment).
- Reached ~95% near-expert quality within 3 cycles (<5 s latency).
- 98.8% precision via threshold tuning, class balance, and ensemble critics.
Mid-Day Meal Analysis using ML (Publication)
Jan 2023
Public PolicyRegressionData Analysis
- Built a state-wise data pipeline; cleaned, normalized, and imputed missing nutrition fields.
- Ran logistic/linear regression to quantify disparities and highlight high-variance states.
- Produced policy-facing visual summaries and age-group recommendations.
Alzheimer’s Disease Severity Detection (Publication)
Aug 2022
HealthcareRandom ForestFeature Engineering
- Engineered features from clinical assessments; compared tree ensembles vs baselines.
- Random Forest + tuned hyperparameters; reported stratified CV metrics.
- Explained top drivers (e.g., cognitive scores) with importance analysis.
Brain Tumor Classification (Publication)
Dec 2021
Medical ImagingKNNML Research
- Extracted MRI features; optimized KNN (k, distance metric) for multi-class tumors.
- Achieved ~97.6–99% accuracy across glioma, meningioma, and pituitary classes.
- Presented at World Conf. on Computational Intelligence; reproducible notebooks.
Automatic Rain-Shade Simulation (Arduino + Proteus)
Jun 2020
EmbeddedSimulationAutomation
- Simulated an UNO + rain sensor controlling dual servos for a retractable roof.
- Implemented control logic with hysteresis to avoid flicker on intermittent rain.
- Documented circuit, algorithm, and test cases for home-automation use.
Education
Stony Brook University
M.S. in Data Science
Aug 2024 – May 2026 (Expected) • New York, USA
Relevant Coursework
Introduction to ProbabilityData Analysis INatural Language ProcessingStatistical Learning (R)Statistical Computing (R)Large Language ModelsData Management
Vishwakarma Institute of Technology
B.Tech in Information Technology (CGPA: 8.97/10)
Aug 2019 – May 2023 • Pune, India
Leadership & Volunteering
Highlights
MentorshipCode ReviewsWorkshop FacilitationCommunity Outreach
I enjoy leading with clarity and empathy—codifying processes, unblocking people quickly, and turning workshops into practical takeaways.
Microsoft Learn Student Club
Android Development Lead • Aug 2021 – Aug 2022
- Led the Android dev pod; ran stand-ups and reviews.
- Hosted hands-on workshops and mentored juniors.
Uttkarsh – Social Initiative
Volunteer • Oct 2021
- Made short videos teaching the 'Symmetry' topic and submitted them to the program.