Skip to content
Krish Nalam

Krish Nalam

Software & ML Engineer

Coffee · Code · Concentration

UofT Webring

About

I’m a Computer Science student at the University of Toronto who likes turning messy, large-scale problems into systems that quietly just work.

Across six internships at RBC, RBC Borealis, and a fintech startup, I’ve shipped RAG pipelines, CI/CD and certificate automation, ETL services, and ML forecasting models. I’m also on aUToronto, U of T’s self-driving car team. On weekends I’m usually at a hackathon, debugging life’s challenges with a dash of humour and caffeine.

Currently studying Honours B.Sc. Computer Science (Co-op) at University of Toronto (expected April 2029).

Tech stack: Python, TypeScript, Java, C/C++, Rust, SQL, Bash, PyTorch, TensorFlow, Hugging Face, LangGraph, RAG, OpenCV, React, Next.js, React Native, Node.js, FastAPI, Spring Boot, Linux, Docker, Kubernetes, AWS, Ansible, GitHub Actions, PostgreSQL.

Experience

  1. RBC BorealisSep 2026 — Present

    Machine Learning Engineer Intern · RBC Borealis

    • Python
    • RAG
    • FastAPI
    • Vector Search
    • Docker
    • Kubernetes
  2. aUTorontoSep 2026 — Present

    Software Developer · aUToronto

    • Python
    • C++
    • Linux
  3. RBCMay 2026 — Aug 2026

    Software Engineer Intern · RBC

    • Built an end-to-end CI/CD pipeline on Red Hat Ansible deploying an auth service to Windows Server across 4 environments, cutting deploy time 85% and eliminating 30+ manual steps.
    • Automated the TLS certificate lifecycle against an OAuth2 PKI CA, cutting provisioning from 2 days to 5 minutes using in-memory PKCS12 bundles and HashiCorp Vault so no key touches disk.
    • Ansible
    • Python
    • HashiCorp Vault
    • OAuth2
    • CI/CD
  4. RBCMay 2025 — Aug 2025

    Software Developer Intern · RBC

    • Engineered a Java/Spring Boot ETL service consolidating HSBC data post-merger, reformatting 30GB+ of weekly data for 12+ mainframe jobs and migrating SOAP endpoints to REST.
    • Wrote PostgreSQL stored procedures archiving transaction history (database size down 15%) and modernized legacy RBC Express code to ES6+, contributing $100K in annual savings.
    • Java
    • Spring Boot
    • PostgreSQL
    • REST
    • ES6

Projects

  1. Project Tantalus

    Coming soon
  2. Lens

    An agentic research assistant that plans multi-step research and cites verified sources.

    • LangGraph agent that plans multi-step research, retrieves from uploaded documents and web search via Chroma, and drafts a cited report.
    • Fine-tuned DeBERTa-v3 on FEVER to flag unsupported claims, serving per-claim confidence via FastAPI and a streaming Next.js UI, with retrieval scored by RAGAS.
    • LangGraph
    • FastAPI
    • Next.js
    • Chroma
    • DeBERTa-v3
    • RAGAS
  3. A live community platform showcasing UofT students’ portfolios in one connected network.

    • Locked member data behind Postgres row-level security with a column allowlist keeping emails out of public payloads; verified site ownership via an HMAC-SHA256 DNS TXT check.
    • Moved avatar processing into the browser (crop, 512px, AVIF) and uploaded direct to S3 via presigned URLs with immutable CloudFront caching, cutting image payload 80%.
    • Next.js
    • TypeScript
    • Supabase
    • AWS S3
    • CloudFront
  4. FITRIS

    Winner @ JamHacks 8

    An interactive fitness game that maps real exercises to Tetris controls via computer vision.

    • Mapped 5 exercises to Tetris controls in real time, classifying a 33-point MediaPipe skeleton with a state machine over 12+ joint angles that counts only full reps.
    • Architected a 3-process pipeline (OpenCV, Flask, React) replacing file polling with Server-Sent Events and a monotonic sequence counter for exactly-once delivery.
    • Python
    • OpenCV
    • MediaPipe
    • Flask
    • React
  5. DevDuels

    Winner @ HackTheValley

    A web game that encourages effective code reviews through competition and real-time AI feedback.

    • MongoDB + GitHub API back end for auth, storage, and user integration.
    • Context-aware feedback via embeddings & RAG with LangChain, semantically analyzing code.
    • Next.js
    • MongoDB
    • LangChain
    • Tailwind

Logbook

Coming soon.