Mayank Goley

01. ABOUT ME

Hey, I'm Mayank Goley. Glad you stopped by.

I'm a backend engineer pivoting hard into applied AI. Over the last six years my code has processed two million daily meter readings for Ontario's electric grid, forecast demand across PepsiCo's supply chain, and shipped TeachWise, a tutoring marketplace I run with a business partner. I'm finishing my Master's in Computer Science at Cal State Fullerton and looking for what comes next.

Most of what I do these days is build. HireNvy is my CSUF capstone, an AI career intelligence platform where I trained a custom DistilBERT classifier that hit 97.3% macro F1 on resume parsing. Financial Intelligence System is a multi agent RAG project on SEC filings with LangGraph and hybrid Qdrant + BM25 retrieval. Ask is a desktop offline first RAG app I'm shipping in Tauri and Rust. Alongside those, ML builds across lane detection (ResNet18), a traffic sign CNN, face mask detection, speech emotion recognition, IMDB sentiment, hourly energy forecasting (LSTM and Transformer), and house price regression. TeachWise rounds it out, a live full stack tutoring marketplace I run with a business partner, built on Flask, Next.js, PostgreSQL, and Stripe Connect.

I graduate in 2026 and I'm hunting for SDE, ML Engineer, or AI Engineer roles at teams that ship and treat AI as a craft rather than a vibe.

When I'm not coding you'll find me watching MLB games, cooking for friends, hitting the Orange County beaches, or at LA tech meetups.

02. WORK EXPERIENCE

Machine Learning Engineer

Nov 2022 – Jun 2023

Infosys (Client: PepsiCo — Demand Forecasting Platform) | Chandigarh, India

Created PySpark data transformation modules across 8 regional SAP configurations, handling locale-specific date formats, currency conversions, tax structures, and SKU hierarchy mappings with full test coverage.
Wrote SQL pipelines joining SAP sales (VBAK/VBAP), inventory (MARD), and material (MARC) tables with promotional calendars, converting base units to business aggregates across SKU-location combinations.
Implemented an outlier detection and classification system using IQR flagging with rule-based routing against trade calendars, integrating a planner review queue back into the ML training pipeline.

Software Developer

Feb 2020 – Jul 2022

Capgemini (Client: Hydro One — Grid Analytics & Energy Platform) | Pune, India

Developed Java batch pipelines (Spring Batch) ingesting 2M+ daily smart meter readings via SFTP, with timestamp validation and anomaly detection, reducing processing failures by 40%.
Optimized Spring Boot REST endpoints with JPA query tuning and response caching, cutting API response time by 35% for the customer-facing consumption and outage portal.
Maintained cross-system integration pipelines between GIS, OMS, CIS, and MDM using SOAP services and SFTP, resolving 95%+ of data sync failures within SLA.
Rewrote compliance reporting SQL with indexing and date-range partitioning, reducing NERC/PUC report generation from 2 hours to 40 minutes.

03. PROJECTS

Energy Forecasting

Deep learning model forecasting 24 hours of energy demand on 16 years of hourly PJM grid data. Built with PyTorch, comparing LSTM and Transformer architectures.

PyTorch, LSTM, Transformer

Traffic Sign Classification

Multi class classifier for German traffic signs across 43 categories. Built with PyTorch, comparing a scratch CNN against a frozen ImageNet pretrained ResNet18.

PyTorch, CNN, ResNet18

Financial Intelligence System

Multi agent RAG system that analyzes SEC filings with Bull, Bear, Synthesis, and Verifier agents debating the same query. Built with LangGraph, FastAPI, React, and a hybrid Qdrant plus BM25 retrieval pipeline using Reciprocal Rank Fusion.

LangGraph, FastAPI, React, Qdrant, BM25

Speech Emotion Recognition

Eight class speech emotion classifier on the RAVDESS dataset using log mel spectrograms. Built with PyTorch and librosa.

PyTorch, librosa

TeachWise

A full stack tutoring marketplace connecting students with tutors. Features real time video sessions, geo based search, and split payment processing. Built with Flask, Next.js 14, PostgreSQL with PostGIS, Stripe Connect, and Jitsi.

Flask, Next.js 14, PostgreSQL, PostGIS, Stripe Connect, Jitsi

HireNvy

AI career intelligence platform that parses job descriptions, scores resumes, and matches candidates to roles. Built with Django, FastAPI, Next.js, a custom DistilBERT classifier, Sentence-BERT, and Qdrant.

Django, FastAPI, Next.js, DistilBERT, Sentence-BERT, Qdrant