Ishika Tomar
Replacing manual data entry and physical inventory counts wi
New to Contra
Ishika is building their profile!
Overview:
Architected an enterprise-grade command dashboard and operations center designed to streamline support ticket telemetry, severity clustering, and operational triage workflows in real time.
What I Did:
Built an interactive command center interface to help internal support teams track and resolve ticket backlogs instantly.
Implemented real-time severity clustering and channel entry metrics to prioritize critical operational roadblocks.
Streamlined workflow telemetry to give engineering and support leadership clear visibility into support health and response metrics.
Tech Stack: Python, Streamlit, Data Analytics, Telemetry, REST APIs
0
2
Gateway Command Center — Real-Time Payment Operations Dashboard
A monitoring and incident-response system for payment infrastructure, built to catch and resolve gateway failures before they become revenue loss.
What it does: Streams live transaction data, tracks failure rate and revenue-at-risk in real time, monitors multi-region edge cluster health (latency, status), and automatically triggers failover + incident triage when a region degrades. Includes a manual fault-injection simulator for stress-testing failure scenarios (rate limits, insufficient funds, fraud checks).
Key features: Live telemetry stream analytics, automated incident triage queue with action plans, multi-region failover, and a RAG-based semantic search over internal SLA/compliance policy documents with confidence scoring.
Stack: Python, real-time streaming, RAG pipeline.
Available for freelance work in fintech ops tooling, real-time systems, and RAG-based internal search
0
8
OmniPulse — Enterprise Churn Prediction & Prevention Engine
An analytics system that ingests customer feedback and account data, uses NLP to cluster the true root causes of churn, and automatically triggers retention playbooks for at-risk accounts.
What it does: Scores every client account by churn risk (Critical/High/Medium/Low), calculates MRR/revenue at risk, tracks sentiment trends over time, and surfaces the top churn catalysts ranked by frequency — so teams can fix the highest-impact problem first instead of reacting account-by-account.
Key features: NLP-based root-cause clustering, LSTM sequence memory for trend tracking, per-account risk cards with Stripe billing status and last-contact data, auto-generated retention playbooks with execution tracking, and a persistent data workspace with a raw-data inspector.
Stack: Python, NLP/clustering, LSTM.
Available for freelance work in churn analytics, customer success tooling, and applied NLP.
0
11
VisiTrack Enterprise — Privacy-Compliant Computer Vision Analytics
A real-time visitor tracking and heatmap analytics dashboard built for retail/venue use cases where privacy compliance (GDPR/CCPA-style requirements) is non-negotiable.
What it does: Detects and tracks people from a video feed to generate footfall velocity, occupancy, and zone-level density heatmaps — without persisting any identifiable image data. Each detected individual is assigned a cryptographic hash instead of stored biometric/video data.
Key features: Zero-persistence cryptographic scrubbing, multiple overlay modes (heatmap, bounding box + hash, skeleton tracking), adjustable detection confidence, live telemetry charts, and a built-in compliance audit log with JSON payload inspection and CSV export.
Stack: Python, OpenCV, real-time video processing.
Available for freelance builds in computer vision, retail analytics, and privacy-first tracking systems.
0
16