Simranjeet S.

Proof of Experience

Adeptmind · Jan 2023 – Present

• Built a GenAI-powered QA evaluation platform with Python, Django, Mirascope, LiteLLM, PostgreSQL, and Celery, automating customer-call auditing and increasing QA coverage from 8% to 100%.
• Engineered low-latency, event-driven AI workflows using WebSockets, structured LLM streaming, and asynchronous processing, reducing end-to-end system latency by 10%.
• Designed state-aware conversational QA pipelines with Django models and Jinja interfaces to validate compliance workflows, including conditional disclaimer verification before payment processing.
• Reduced infrastructure costs by 40%+ through monolithic deployment optimization, Celery worker offloading on spot instances, and efficient orchestration of CPU-intensive embedding workloads.
• Developed a multi-accent voice training agent with the LiveKit Agents framework to improve conversational-AI training across diverse speech patterns.
• Automated staging deployments with Docker Compose, GitHub Actions, and CI/CD workflows while strengthening API security and improving inventory-management efficiency.
• Improved large-scale data-gathering pipelines with Selenium, cURL, and vector-database integrations to support chatbot intelligence and product-recommendation workflows.
• Added product-page deep linking to improve navigation and collection of user-behavior signals.

Anakin (YC S21) · Jul 2022 – Jan 2023

• Built a multimodal price-matching platform with Python, Django, OpenCV, and PostgreSQL to identify identical products across e-commerce platforms.
• Engineered OCR and image-difference pipelines with OpenCV to normalize fragmented catalog data and improve product-matching accuracy.
• Developed scalable ingestion pipelines and analytics dashboards with BigQuery and Looker Studio, enabling brands to monitor pricing and inventory across marketplaces.
• Built a Django-based annotator-management system for multi-user product verification and dataset correction workflows.
• Optimized cloud infrastructure and backend services, reducing operational costs by more than 30% while improving pipeline scalability.
• Reverse engineered web and Android applications to discover internal APIs and automate competitive-product intelligence collection.

Sudoviz · Jun 2021 – May 2022

• Progressed from Security Engineering Intern into a full-time software-engineering role on the founding team.
• Built an Application Security Posture Management platform with FastAPI, Docker, Kubernetes, Neo4j, and AWS to centralize SAST and SCA vulnerability-analysis workflows.
• Developed isolated containerized scanning services that orchestrated multiple security scanners and stored artifacts in Amazon S3 for scalable vulnerability management.
• Built reusable React components and integrated them with FastAPI services to deliver software-composition and vulnerability insights.
• Engineered services to parse and validate 7+ SBOM formats, enabling identification of compromised packages and dependency vulnerabilities with high data freshness.
• Automated CI/CD with GitHub Actions and Kubernetes deployment pipelines to improve release reliability and infrastructure scalability.
• Worked on container security and source-code vulnerability management; deployed containers to Google Cloud Run with API Gateway and automated GitHub-to-Firestore updates.

MIDAS: Multimodal Digital Media Analysis Lab · May 2020 – Apr 2021

• Developed a Devanagari information-retrieval system for ASHA healthcare workers using PyTorch, Flask, Docker, and AWS to improve access to maternal and child-health information in rural India.
• Applied Siamese neural networks and transfer learning to improve semantic retrieval for Hindi healthcare queries.
• Built and deployed deep-learning inference services in Docker on AWS, with secure cloud-to-edge communication.
• Engineered GSM-based SMS relay automation on Linux edge devices, reducing messaging costs by nearly 700x compared with traditional SMS providers.
• Co-authored “Hindi Chatbot for Supporting Maternal and Child Health Related Queries in Rural India,” published at ClinicalNLP 2023.

NCIIPC India (A unit of NTRO) · Sep 2020 – Dec 2020

• Researched classification methods for network-intrusion detection systems.
• Built a two-stage attack-classification pipeline using a binary classifier followed by a multiclass classifier.
• Delivered a working prototype and technical report documenting results and improvement opportunities.

CERT-In · Dec 2019 – Jan 2020

The Indian Computer Emergency Response Team (CERT) is an office within the Ministry of Electronics and Information Technology of the Government of India.
It is the nodal agency to deal with cyber security threats like hacking and phishing.
• Summarised the survey papers between 2005 to 2019 and wrote a review paper on it
• Orchestrated literature survey on inter-domain routing protocol vulnerabilities & mitigation techniques

Indian Army. · Jun 2019 – Jul 2019

The Military College of Telecommunication Engineering is the engineering training establishment for the Corps of Signals, of the Indian Army.
Got the chance to work under Lt.
Colonel Bhatt to address the problem domain of object detection using machine learning.
• Orchestrated high quality data for the Military Object Detection Project at MCTE
• Dataset of about 500 images was collected which was then further augmented

No completed sessions yet. Anything Simranjeet finishes shows up here with the evidence behind it.

Netaji Subhas University of Technology · 2018 — 2022