Divit Mallah
Open to work1 years experienceAI Engineer
AI Engineer specializing in LLMs, backend systems, and data analytics.
Standing Rank
Rank Not Available
Developer Badges
No badges earned yet
Skills & Technologies
41 skillsGitHub Activity Evidence
GitHub Footprint Not Connected
Recruiters value verified code contributions. Linking a GitHub account displays live heatmaps, active contribution metrics, repository languages, and push activity tags.
Work Experience
AI Engineer
Naptick
- Built production NestJS and TypeScript services powering AI analytics, wearable ingestion, and user-facing features.
- Developed LangChain workflows using Gemini, structured outputs, and health-data tools for personalised sleep insights.
- Integrated the Gemini Files API to ground biometric explanations in sleep, nutrition, and recovery research.
- Generated personalised insights connecting food intake and biometric changes with evidence-backed recommendations.
- Designed Apple Health and Oura pipelines converting wearable data into sleep and recovery metrics.
- Built Flutter and iOS features by integrating APIs, developing UI screens, and implementing app-lock workflows.
- Improved reliability through caching, cron jobs, MongoDB optimisation, and PM2 deployment monitoring.
- Contributed to Raspberry Pi sensor and audio pipelines supporting device automation and IoT communication.
Data Scientist Intern
HERE Technologies
- Built Python and DuckDB pipelines querying 100M+ mobility records faster than Pandas workflows.
- Applied HDBSCAN to GPS probe trajectories, identifying mobility clusters and traffic patterns across cities.
- Used OpenCV rules to detect and validate 150+ satellite roundabouts with 90% precision.
- Built hotspot detection using density curves, interpolation, and temporal anomaly analysis on city-scale datasets.
- Created ESRI GA heatmaps and geospatial visualizations for multi-city trend tracking and reporting.
Project Trainee
Bhabha Atomic Research Centre
- Studied Sysmon for Windows endpoint security monitoring and event-log based threat visibility.
- Configured Winlogbeat, Logstash, and Elasticsearch to stream endpoint events to a centralized logging pipeline.
- Used Elasticsearch APIs and Kibana dashboards to analyze endpoint activity and visualize security events.
Web Development Intern
Smartlink Holdings Ltd.
- Built responsive frontend modules using React and Bootstrap for a modular ERP platform.
- Developed reusable React components, improving code reusability and reducing repeated frontend development effort.
- Implemented cross-device responsive layouts for desktop and mobile usage.
Projects
Production Sleep Intelligence Platform
- Built sleep-tech backend systems integrating wearable data, ML predictions, and AI-generated user insights.
- Designed pipelines processing sleep, HRV, heart rate, respiration, steps, and activity metrics.
- Implemented sleep quality and recovery workflows using rolling metrics and time-series health signals.
- Built LLM explanation pipelines generating personalized sleep insights from biometric and behavioural data.
Cricket Analytics Agent with LLM + DuckDB + LangChain
- Built a Gemini-based SQL agent using LangChain and DuckDB to query IPL datasets from 2008–2024.
- Enabled natural language to SQL conversion using schema-aware prompts and fuzzy player-name resolution.
- Implemented playoff-stage inference to support ambiguous cricket queries with missing match or stage context.
- Deployed an interactive Chainlit UI for live IPL analytics and user exploration.
Federated Learning for Medical and Tabular Classification
- Built a federated learning pipeline for medical image and tabular classification across IID and Non-IID client partitions.
- Implemented secure gradient aggregation and benchmarked multiple federated optimization strategies.
- Achieved 95.5% accuracy under IID settings and analyzed degradation under Dirichlet Non-IID data skew.
- Refined tree-merging logic to reduce global model loss across heterogeneous client distributions.
Flood Mapping in India using SAR Imagery and Deep Learning
- Curated pre- and post-flood Sentinel-1 SAR imagery across multiple Indian states using VH gamma0 data.
- Processed imagery with ESA SNAP and Google Earth Engine , generating flood masks using OTSU and K-means.
- Trained a custom U-Net model, achieving 95.6% accuracy, 76.4% IoU, and 86.6% F1-score .
- Delivered a scalable flood-detection pipeline for rapid disaster response and remote-sensing analysis.
Leaderboard Standings
Leaderboard Position Pending
Global test scores, peer standing percentiles, and algorithm leaderboard ranks are updated dynamically.
Assessment Highlights
Assessments Not Completed
Coding evaluations, system assessment results, and conceptual score badges will appear here after taking a test.
AI Collaboration Score
AI Collaboration Score Pending
Developer coding behavior, assistant cooperation, and AI pair-programming indicators are evaluated during live coding sessions.
Role Compatibility Profile
Role Compatibility Analysis Pending
Custom matching reports, candidate role compatibility percentiles, and core engineer strength profiles are processed once conceptual code screenings are complete.
Achievements
INSPIRE Scholarship
Recipient of the INSPIRE Scholarship by the Department of Science and Technology.
Top 1% Nationwide Ranking
Ranked in the top 1% nationwide in 12th Board Examinations, Science stream.
About Details
Professional Bio
AI Engineer with expertise in building production-grade NestJS services, LLM workflows, and data-driven applications. Experienced in integrating wearable health data, geospatial analytics, and federated learning systems.
M.Sc. (Hons.), Mathematics and B.E. (Hons.), Computer Science
Birla Institute of Technology and Science, Pilani - Pilani Campus (2020 - 2025)