Project Showcase | Harsh Gidwani

01

Featured

GrowAI

End-to-End AI Marketing Automation Platform

GrowAI is a fully-automated marketing intelligence platform engineered to convert raw data into multi-channel marketing assets. The system ingests product data via CSV/Excel, preprocesses it using rule-driven attribute extraction, and generates content using a fine-tuned LLM pipeline built on top of the Llama architecture. GrowAI produces commercial-grade ad scripts, short-form video content plans, email campaigns, and social media copy, adapting style through transformer-based tone modulation.

The platform incorporates a scalable event-driven backend architecture, GPU-optimized inference endpoints, and real-time performance feedback loops. By combining automated content generation with conversion-focused analytics, GrowAI minimizes manual creative effort and delivers a measurable uplift in marketing ROI without requiring a dedicated creative team.

10k+ Assets Generated

5x ROI Increase

80% Time Saved

Multi-Channel Distribution

Technology Stack

Llama Architecture Python Transformers GPU Optimization FastAPI Event-Driven Backend NLP Fine-Tuned LLM

Key Features

✓

Automated CSV/Excel data ingestion with rule-driven attribute extraction

✓

Fine-tuned Llama-based LLM for commercial-grade content generation

✓

Multi-channel asset creation: ad scripts, videos, emails, social media

✓

GPU-optimized inference endpoints for real-time performance

✓

Transformer-based tone modulation and style adaptation

✓

Conversion-focused analytics with performance feedback loops

Request Details

💼 For detailed project information, technical documentation, or collaboration opportunities, please contact me at harshgidwani2007@gmail.com

02

Live

Finura

Smart Financial Planning & Budget Intelligence System

Finura is a personal finance automation engine built to analyze transaction behavior, predict spending anomalies, and enforce user-defined financial objectives. It leverages a hybrid inference system: rule-based categorization powered by deterministic spending logic, combined with LSTM-driven recurrent forecasting for expense prediction and financial health scoring.

The platform dynamically recommends budgets using statistical models fused with sequence-based forecasting, and visually communicates spending behavior through an adaptive analytics dashboard. Security and privacy are central to the system architecture, employing AES-256 encrypted storage, compartmentalized process execution, and hashed identity mapping. Finura transforms scattered financial data into actionable insights, enabling users to achieve long-term financial control through AI-augmented planning and automated, data-driven decision support.

95% Prediction Accuracy

AES-256 Encryption

Real-Time Analytics

LSTM Forecasting

Technology Stack

LSTM Networks Python TensorFlow AES-256 Encryption Time Series Analysis Statistical Modeling Data Visualization Secure Architecture

Key Features

✓

Hybrid inference: rule-based + LSTM recurrent forecasting

✓

Spending anomaly detection with predictive alerts

✓

Dynamic budget recommendations using statistical fusion models

✓

AES-256 encrypted storage with compartmentalized execution

✓

Adaptive analytics dashboard with behavioral insights

✓

Financial health scoring with actionable recommendations

Request Details

💼 For detailed project information, technical documentation, or collaboration opportunities, please contact me at harshgidwani2007@gmail.com

03

Enterprise

CognisCRM

AI-Driven Customer Lifecycle & Business Intelligence Suite

CognisCRM is a next-generation customer lifecycle platform that applies machine intelligence to lead acquisition, retention optimization, and business analytics. Built on Next.js for reactive UI orchestration and integrated with Llama 3.1 for multimodal lead intelligence, the system autonomously prioritizes customer pipelines using predictive scoring models.

It segments prospects through unsupervised clustering, identifies churn risk using behavioral profiling, and enriches CRM operations with automated communication, recommendation routing, and opportunity forecasting. Enterprise scalability is achieved using modular API endpoints, microservice-aligned backend services, and a plugin-ready AI inference layer. By merging data analytics, generative intelligence, and operational automation, CognisCRM delivers a measurable reduction in lead leakage, faster deal closures, and a strategically optimized sales workflow without requiring manual intervention.