AI/ML Engineering Leader & Full-Stack Architect

HARSHIT SONI New York, NY Linkedin (harshitsoni01) harshitsoni.com GitHub (harshitsoni01) PROFESSIONAL EXPERIENCE NextLM Inc. United States, Remote Head of Engineering, Founding Team Aug 2026 – Present ● Lead engineering and 0-to-1 product development for NextLM’s AI platform, owning full-stack architecture and hands-on coding across React, TypeScript, and Python, scaling to 100+ customers, including enterprise accounts with custom CRM integrations. ● Architect large-scale AI and data infrastructure processing 35B+ rows daily across 350M+ profiles, powering prospecting, enrichment, and buyer-intent intelligence at scale ● Productionize and scale Savant, NextLM’s proprietary prospecting model, outperforming frontier models on internal benchmarks and deployed on NVIDIA A100 GPUs with LoRA adapters ● Owned cloud architecture across GCP, APIs, microservices, CI/CD, Docker, and Kubernetes, working directly with Google Cloud and NVIDIA engineers on AI infrastructure and model deployment ● Led engineering efforts to bring NextLM into Gemini Enterprise, advancing the platform’s Google-native architecture and enterprise AI distribution Senior Machine Learning Engineer Jan 2026 – Aug 2026 ● Built and deployed production LLM applications, AI agents, and ML systems powering NextLM’s sales intelligence platform ● Developed scalable backend services, APIs, data pipelines, and enterprise integrations using TypeScript, Python, and GCP ● Shipped customer-facing AI features from prototype to production while improving platform performance, reliability, and scalability Cognitive Network Solutions Inc (CNS), Machine Learning Ops Engineer Ridgefield Park, NJ Aug 2025 - Dec 2025 ● Developed neural network models for RAN performance prediction and cell tower tilt optimization, enhancing network efficiency ● Designed and deployed end-to-end MLOps workflows for 6G wireless optimization graph RL models, reducing model deployment time and improving reliability across production systems ● Built an automated data pipeline processing 100K+ data every 5 min from Nvidia Sionna using FastAPI, MCP endpoints, & GCP, enabling real-time model training. Containerized and deployed microservices with Docker & Kubernetes, ensuring scalable AI system ● Collaborated with C-suite to align AI-driven cloud systems with CNS’s vision for self-optimizing 6G (AI-RAN) networks Julius Education, Machine Learning Engineer New York, NY Jun 2024 – Sept 2024 ● Built ML-powered career dashboards using RAG, prompt engineering, and LLMs to improve renewable-energy career data ● Developed data pipelines, web scraping, and cloud integrations for government datasets and client projects, including NREL Trust Scout, Software Engineering Intern New York, NY Jun 2023 – Aug 2023 ● Built an automated email marketing system that increased client replies by 40% using Python, AWS, Django, SQL, APIs, and AI ● Analyzed 10M real estate transactions and 1.6M agents using ML and statistical modeling to generate business insights TECHNICAL SKILLS Languages: Python, TypeScript, SQL, JavaScript, C++, Java, Shell, HTML/CSS AI/ML: LLMs, AI Agents, RAG, MCP, Machine Learning, Deep Learning, NLP, TensorFlow, PyTorch, Scikit-learn, CUDA Backend & Data: FastAPI, Flask, REST APIs, Microservices, ETL/Data Pipelines, PostgreSQL, Pandas, NumPy Cloud & Infrastructure: GCP, AWS (S3, EC2, IAM), Docker, Kubernetes, CI/CD, Git Engineering: System Architecture, Distributed Systems, Enterprise Integrations, Data Infrastructure, MLOps, Automation EDUCATION Binghamton University, State University of New York Master of Science in Computer Science, Artificial Intelligence Track Cumulative GPA: 3.6 /4.00 RESEARCH EXPERIENCE Binghamton University, Research Assistant – Human-Computer Interaction Jan 2024 – Mar 2024 ● Researched ML model integrating computer vision and natural language processing (NLP) to translate real-time ASL into English text ● Improved hand recognition accuracy using neural networks and image processing for more precise ASL interpretation Binghamton University, Research Assistant – Semiconductor Fabrication Jan 2023 – May 2023 ● Researched ML autoencoder models to develop microscopic filter designs for faster semiconductor manufacturing ● Trained an adversarial model (PyTorch, CUDA) to predict and generate filter shapes based on light transmission across wavelengths IEEE, Published Research Paper Title ‘Fake News Detection Using NLP and Logistic Regression’ Sept 2021 PROJECT EXPERIENCE Zillow Page Summarizer: Chrome Extension Assistant, Independent Project (link) Jul 2024 – Sept 2024 ● Developed an extension that summarizes Zillow property listings, providing real-time insights using web scraping and NLP ● Built the extension’s backend using the Flask framework and integrated an LLM for sentence processing and APIs for data retrieval ● Tech stack: LLM, Python, Flask, Pyzill, API, HTML/CSS, JavaScript, Chrome Extension Framework, Natural Language Processing LEADERSHIP & HONORS Hackathon (Secured 2nd out of the 49 teams) (link)Jul 2020 ● Led a team of three to develop a ‘Fake News Detection Model using Machine Learning and Natural Language Processing’ ● Built an ML model to classify news as "fake" or "real," providing users with verified news and enhancing digital media's reliability